- Email: [email protected]
Magnetic resonance microscopy of spinal cord injury in mouse using a miniaturized implantable RF coil
Journal of Neuroscience Methods 159 (2007) 93–97
Magnetic resonance microscopy of spinal cord injury in mouse using a miniaturized implantable RF coil Mehmet Bilgen ∗ Hoglund Brain Imaging Center and Department of Molecular and Integrative Physiology, The University of Kansas Medical School, Mail Stop 3043, 3901 Rainbow Blvd, Kansas City, KS 66160, United States Received 27 March 2006; received in revised form 26 May 2006; accepted 27 June 2006
Abstract A magnetic resonance neuroimaging method is described for high-resolution imaging of spinal cord injury in live mouse. The method is based on a specially designed radio frequency coil system formed by a combination of an implantable coil and an external volume coil. The implantable coil is a 5 mm × 10 mm rectangular design with a 9.1 pF capacitor and 22 gauge copper wire and optimal for surgical implantation over the cervical or thoracic spine. The external volume coil is a standard birdcage resonator. The coils are inductively overcoupled for imaging the spinal cord at 9.4 T magnetic field strength. The inductive overcoupling provides flexibility in tuning the resonant frequency and matching the impedance of the implanted coil remotely using the tuning and matching capabilities of the volume coil. After describing the implementation of the imaging setup, in vivo data are gathered to demonstrate the imaging performance of the coil system and the feasibility of performing MR microscopy on injured mouse spinal cord. © 2006 Elsevier B.V. All rights reserved. Keywords: Magnetic resonance imaging; Microscopy; Inductive overcoupling; Radio frequency coil; Implantable coil; Spinal cord
1. Introduction Rodent models are the backbone of neuroscience research in experimental spinal cord injury (SCI) (Kwon et al., 2002, and the references therein). According to the current trend, however, the focus is shifting from rat to paraplegic murine models because of the availability of genetically engineered transgenic and mutant mouse strains along with molecular tools (Guertin, 2005; Rosenzweig and McDonald, 2004; Steward et al., 1999). The growing number of such efforts is likely to produce new targets for therapeutic strategies aimed at promoting neuronal protection, reorganization and functional recovery following SCI in humans (Bloom et al., 2005). Magnetic resonance imaging (MRI) modalities have long been used to remotely probe the spinal cord (SC) in rodent models, and radio frequency (RF) coil systems were developed with a special feature that allowed the imaging coil to be surgically implanted at the level of the SCI in rats (Bilgen et al., 2001; Bilgen, 2004; Ford et al., 1994; Silver et al., 2001; Wirth et al., 1993). The implantable coils offer improved signal-to∗
Tel.: +1 913 588 9079; fax: +1 913 588 9071. E-mail address: [email protected].
0165-0270/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jneumeth.2006.06.024
noise ratio and small field-of-view selection due to the small dimensions of the coil and compact footprint of the RF excitation/signal detection. Both of these features aid in the acquisition of high-resolution images from the tissue of interest only. In our studies, this arrangement provides a practical solution for spatially localized imaging of the SCI and enhances the quality of the MR data acquired in longitudinal studies. Signal transmission and reception with the implanted coil is achieved via mutual-inductive coupling with a larger coil placed external to the body. We have shown that strengthening the mutual coupling increases the functionality of the implantable coil (Bilgen, 2006). Inductive overcoupling is the condition of electromagnetic interaction between two coils beyond the point at which the resonance peak of the response curve for the current in the coupled coil splits. This allows the frequency response characteristics of the implantable coil with fixed electrical elements to be modified as needed via an external coil with tuning and matching capacity. Such ability provides an environment for reliable and repeatable acquisitions of MR data, which is a highly desirable feature for studying the progression of the pathology in injured SC over time on the same animal. Previously, we have shown that an implantable coil developed for imaging rat SC can also serve as a surface coil to
94
M. Bilgen / Journal of Neuroscience Methods 159 (2007) 93–97
Fig. 1. In vivo sagittal image of a mouse spinal cord injured (arrow) at the thoracic level T11. The image was acquired using an inductively coupled surface coil developed originally as an implantable coil for imaging rat SCI. The acquisition parameters were TR /TE = 2500/12 ms, FOV=30 mm X 15 mm, image matrix = 256 X 128, slice thickness = 0.5 mm and NEX = 2. Fig. 3. Implantable coil and standard quadrature volume coil (4 cm ID) used in these studies. Note that only one channel of the volume coil is used for imaging and the other channel was terminated with 50 ohm load. The arrows point to the 9.1 pf capacitor. A dime was also included in the picture for the purpose of providing relative comparison between the sizes of the coils.
Fig. 2. Sagittal view of the mouse spinal cord at the thoracic and cervical levels. The image was acquired using a quadrature volume coil with 4 cm ID, shown in Fig. 3. The acquisition parameters were TR /TE = 4000/12 ms, FOV = 60 mm X 24 mm, image matrix = 256 X 128, slice thickness = 0.5 mm and NEX = 2.
image normal (Bilgen et al., 2005; Bilgen and Al-Hafez, 2006) and pathological SC (Bilgen, 2004) in mouse. As shown in Fig. 1, this configuration exhibits a region of high sensitivity over the thoracic levels from T10 to T12, and therefore is capable of producing excellent MR data, but only from this spatiallylimited region of the spine. However, in experimental studies, the cord is typically injured at mid-thoracic (Farooque et al., 2006; Ghasemlou et al., 2005) or cervical levels (Anderson et al., 2004). The cord at these levels can be visualized using a volume coil as shown in Fig. 2, but at lower resolution than is possible with the implantable coil and with compromised image quality. In the present study, we investigate the potential of using an implantable coil to improve the visualization of the mouse SC, as has been done in the previous rat studies. To accomplish this task, we introduce a miniaturized implantable coil for experimental mouse SCI studies at 9.4 T magnetic field strength. In the following, we describe the design and components of this coil in detail, and show its performance with data acquired from injured cords in vivo when it is inductively overcoupled to a volume coil with tuning and matching circuitry. We also discuss the benefits and limitations of this coil system in the context of tradeoffs in longitudinal studies. 2. Methods 2.1. RF coil setup We fabricated an implantable coil using a piece of circular (22 gauge, CDA 102) soft bare copper wire (MWS Wire
Industries, Westlate Village, CA) that was bent to form a rectangular “pitched rooftop” shape with dimensions 5 mm × 10 mm. A small 9.1 pF ceramic chip capacitor (American Technical Ceramics, Huntington Station, NY) was soldered at the ends of the wire to complete the circuit, as shown in Fig. 3. The coil was coated with biologically inert silicone elastomer MDX44210 (Factor II, Inc., Lakeside, AZ) for electrical isolation from its surroundings when implanted. The frequency response of the implantable coil was measured using a rectangular (25 mm × 40 mm) pickup coil and a frequency sweeper (Morris Instruments, Inc., Ottawa, Ontario, Canada). The pickup coil was tuned at about 1 GHz and therefore, did not affect the frequency response characteristics of the implantable coil. The coated coil was then sterilized by immersion in Cidex disinfectant for 24 h, followed by immersion in distilled water for 2 h before it was surgically implanted into the animal. For signal transmission/reception with the implantable coil, we employed an inductively overcoupled configuration with a standard quadrature volume coil (Fig. 3), similar to that described previously in rat studies, but the volume coil was for mouse and had a smaller inner diameter of 4 cm (Bilgen, 2006). We terminated one channel of the coil to make it operate as a linear coil. The combination of this volume coil with the implantable coil was used to image the mouse SC. 2.2. Animal surgeries for SCI and coil implantation The experiments were performed on 4-month old C57Bl/6 mice weighing between 20 and 25 g under the protocol approved by the University of Kansas Medical Center Institutional Animal Care and Use Committee. The animal was initially anesthetized with 5% isoflurane in an induction chamber, and then moved to a surgery table, where anesthesia was maintained with a mixture (1.5% isoflurane, 30% oxygen and air) delivered through a nose mask. A midline incision was made from T5 to T12 and the underlying tissue and muscle were dissected. Laminectomy was performed at the level T9. The exposed SC was moderately injured using a 0.5 mm diameter circular-tip injury bit attached to the injury device described previously (Bilgen, 2005). Other
M. Bilgen / Journal of Neuroscience Methods 159 (2007) 93–97
biomechanical parameters for inflicting the injury were set as velocity = 0.75 m/s, displacement = 0.5 mm and compression duration = 100 ms. Next, the implantable coil was centered at the injury site and held in place temporarily to assess which spinous processes needed to be removed to create enough room to stabilize the coil on the rib cage. Once the coil was placed properly on the rib cage, the overlaying paraspinal muscles were sutured at the midline to permanently hold the coil in place and the skin was closed tightly. 2.3. Magnetic resonance imaging MRI scans were performed on a 9.4 T horizontal bore scanner (Varian, Inc., Palo Alto, CA) with 31-cm bore and 400 mT/m gradient coil set. After the surgeries, the animal with the implanted coil was put in the supine position on a plastic cylinder that was cut in half along its long axis. The cylinder was inserted into the volume coil and the implanted coil was centered. To improve the coupling, the volume coil was rotated slightly with respect to the cylinder until two peaks appeared near the proton resonant frequency of 400 MHz on the sweeper’s display. The tuning and matching rods of the volume coil were then engaged to further improve the impedance matching and frequency tuning at resonance. The Plexiglas cradle, supporting the animal and coil assembly, was inserted into the scanner bore, and its placement at the magnet isocenter was confirmed with scout images. Gradient echo (GE) images were acquired in the sagittal and coronal planes to demonstrate the signal patterns associated with the implantable and volume coils over a large field-of-view (FOV). The parameters for these acquisitions were TR /TE = 40 ms/3 ms, flip angle = 45◦ , image matrix = 128 × 128, slice thickness = 2 mm and NEX = 1 and FOV = 35 mm × 50 mm, which resulted in a total acquisition time of 6 s. Next, proton density-weighted images were acquired in high-resolution, multi-slice and interleaved mode in the axial and sagittal, planes using standard spin echo (SE) sequence over smaller FOV. The acquisition parameters for the sagittal images were TR /TE = 4000 ms/12 ms, FOV = 15 mm × 10 mm, image matrix = 256 × 128, slice thickness = 0.5 mm and NEX = 2. The parameters for the axial slices were TR /TE = 4000 ms/12 ms, FOV =6 mm × 6 mm, acquisition matrix = 128 × 128, slice thickness = 1 mm and NEX = 4. The acquisition times for these scans were about 18 and 35 min, respectively. While in the scanner, the mouse was monitored using an MRcompatible small animal monitoring and gating system (Model 1025, SA Instruments, Inc., Stony Brook, NY). This system was also used for respiratory-gated acquisition to minimize the breathing-related image artifacts. The temperature of the animal was kept at 37 ◦ C by circulating warm air with 40% humidity through a 5 cm diameter plastic tube fitted at the back door of the magnet bore. 3. Results and discussion The mouse spine exhibits strong curvature, as shown in Fig. 2. While the cervical and upper thoracic sections of the spine
95
anatomically lie deep in the body, the lower thoracic spine is situated more superficial. In the adult mouse, the cervical spine together with the back of the skull measures about 11 mm in length. Except at its most upper and lower segments, the midthoracic spine remains mostly straight between the levels T5 and T11, measuring about 13 mm in length. The 10 mm length implantable coil in Fig. 3 is sized well for implantation over either the cervical or mid-thoracic segments and can be used for imaging the SC injured at these levels. But, it may not be as easy to implant the coil to image the upper thoracic spine due to the strong curvature at this level. In our experiments, the mice have tolerated the surgical procedures and the SCI at the T9 level. However, due to the small size of the mouse and the challenging environment for surgery, proper implantation of the coil over the tight spinal space required surgical skills. Before the coil was implanted, its frequency response was measured in bench tests in open air using the rectangular pickup coil. The implantable coil was observed to tune at 403 MHz. After implantation into the back of the mouse, the frequency response was measured again using the same pickup coil. This time, the tuning frequency read 398 MHz and the resonant peak was broader. It is typical to see a shift and broadening of the peak in the frequency response of the implantable coil when loaded with tissue impedance after the implantation (Bilgen, 2004). We applied thick elastomer coating on the coil to reduce the dielectric losses by increasing the physical separation between the coil and the surrounding environment. This simple approach made the coil less sensitive to changes in its surroundings and provided stability when it was implanted. When the mouse was placed into the volume coil, the resonance peak was split as expected due to the inductive overcoupling between the volume coil and implantable coil. The first peak was at 387 MHz. The second peak was observed to be closer to the resonance frequency of the scanner. Therefore, this second peak was tuned to 400 MHz by adjusting the matching and tuning circuitries of the volume coil. Fig. 4 shows the whole-body images of the mouse with the coil implanted in the sagittal and coronal views. During transmission, the currents in both the volume and implanted coils produced an excitation field in the body of the mouse. The intensity patterns in the figure depict the signals of both the footprint of the implanted coil and that of the background regions of the mouse body. The homogeneous weak signal in the background was induced by excitation from the volume coil. The inhomogeneous hyperintensity field seen near the site of the implanted coil was formed by both coils, but with a stronger contribution from the implanted coil. Acquisition with a lower FOV around the implanted coil produced the high-resolution images in Figs. 5 and 6. Laminectomized vertebrae as well as the spatial location and extent of the injury along the cord can easily be identified on the thin-sliced sagittal images in Fig. 5. The injured SC tissue is viewed in greater detail at 47 m resolution axial images in Fig. 6. These data clearly demonstrate the capability of the present approach to delineate the gray matter, surrounding white matter and cerebrospinal fluid in normal regions of the cord. More importantly, the injury is clearly depicted as patchy areas of hypointensity, representing hemorrhage in the SC tissue.
96
M. Bilgen / Journal of Neuroscience Methods 159 (2007) 93–97
Fig. 5. In vivo sagittal images of an injured mouse spinal cord acquired with the implantable coil (TR /TE = 4000/12 ms, FOV = 15 mm X 10 mm, acquisition matrix = 256X128, slice thickness = 0.5 mm and NEX = 4). Arrow points to the laminectomy and injury at T9.
Fig. 4. In vivo gradient echo images of a mouse in (a) sagittal and (b) coronal planes acquired with implantable coil that was inductively-overcoupled to the volume coil on the day of surgical implantation. The sagittal image shows the mouse body placed in supine position. The images were acquired at large FOV to demonstrate the signal patterns associated with each coil. The intensities in the images were windowed and scaled to enhance the background signal. Notice that, during transmission, the volume coil generates an excitation field that leads to the hypointense signal (long arrows) seen in the background. The implanted coil generates an excitation field that produces the hyperintense footprint (short arrows). Arrowheads point to the artifacts from heart motion. The acquisition ◦ parameters were TR /TE = 40 ms/3 ms, flip angle = 45 , image matrix = 128 X 128, slice thickness = 2 mm and NEX = 1 and FOV = 35 mm X 50 mm.
Although the inductively overcoupled coil system is capable of providing high quality images, it also poses a limitation because of the physics involved in the MR data acquisition over small FOV (Bilgen, 2006). That is the images acquired inherently contain wraparound signals from tissue outside the selected FOV that has been excited by the volume coil. As shown in Fig. 4, strong signal artifacts caused by motion of the heart are present on the GE images, but signals from other motionless regions are significantly low in intensity. Compared to the GE
data, the SE acquisition at a large FOV produces lower background signal. The consequence of the wraparound is that separating the signals on the regions of interest from the background noise becomes difficult, even after the contrast and brightness of the displayed image are adjusted. Since respiratory gating was employed during our acquisitions, the wraparound artifacts contributing to the SE images in Figs. 5 and 6 were only those from the heart motion, and minimal because of their low signal intensities. In Fig. 7, we provide images of an injured SC acquired ex vivo (i.e., the spine was dissected but the vertebral body was left intact) using the implantable coil that was inductively coupled to the rectangular pickup coil, instead of the volume coil. These images establish the upper boundary on the image quality under the ideal condition of artifact-free acquisition; thereby demonstrating what can ultimately be achievable in terms of imaging the mouse SC. It is interesting to note that the data in Figs. 5 and 6 are comparable to these artifact-free benchmark images in quality, clearly demonstrating the capability of our approach to accomplish the high-resolution imaging task of the SCI in mouse. The approach yields images of adequate quality, spatial sensitivity and specificity, while maintaining good intensity contrast, which makes it easy to differentiate between the injury and the other structures within the SC. Our group as well as others is increasingly using mouse models in experimental SCI studies. However, in vivo MRI of the
Fig. 6. In vivo axial images of the injured mouse in Fig. 5 (TR /TE = 4000/12 ms, FOV = 6 mm X 6 mm, acquisition matrix = 128X128, slice thickness = 1 mm and NEX = 4).
M. Bilgen / Journal of Neuroscience Methods 159 (2007) 93–97
97
Fig. 7. Ex vivo axial images of an injured mouse spinal cord acquired with the implantable coil in Fig. 3 while the vertebral body was still intact (TR /TE = 4000/12 ms, FOV = 4 mm X 6 mm, acquisition matrix = 128X128, slice thickness =0.5 mm, slice gap = 0.5 mm and NEX = 4).
injured mouse SC has received little attention to date (Bilgen et al., 2005; Bilgen and Al-Hafez, 2006; Bonny et al., 2004; Wamil et al., 1998). The previous studies imaged the mouse SC with limited quality and spatial sensitivity by using surface or volume coils. In this paper, we addressed this issue and described an inductively overcoupled coil system consisting of a miniaturized implantable coil. Fabricating this coil was simple and low-cost, and cutting and forming the copper wire and soldering the inexpensive ceramic chip capacitor took less than 15 min. The volume coil was a standard quadrature coil available in any MRI laboratory. The tuning and matching feature of the coil system made it flexible and practical to readjust any changes in the frequency response after its implantation. The results presented above clearly demonstrate the feasibility of acquiring high quality in vivo MR data from the injured SC in mice at 9.4 T using this coil system. Although, the data presented above were anatomical visualization, future studies may take advantage of this coil in using the structural and functional MR modalities, such as diffusion tensor and contrast enhanced imaging modalities, to characterize the longitudinal changes in the pathobiology of the SCI in mice. Acknowledgments The author thanks Dr. Baraa Al-Hafez and Dr. Yong-Yue He for their help in animal preparation, surgeries for coil implantation and MRI scans. References Anderson KD, Abdul M, Steward O. Quantitative assessment of deficits and recovery of forelimb motor function after cervical spinal cord injury in mice. Exp Neurol 2004;190:184–91. Bilgen M, Elshafiey I, Narayana PA. In vivo magnetic resonance microscopy of rat spinal cord at 7 T using implantable RF coils. Magn Reson Med 2001;46:1250–3. Bilgen M. Simple, low-cost multipurpose RF coil for MR microscopy at 9.4 T. Magn Reson Med 2004;52:937–40.
Bilgen M, Al-Hafez B, Berman NE, Festoff BW. Magnetic resonance imaging of mouse spinal cord. Magn Reson Med 2005;54:1226–31. Bilgen M. A new device for experimental modeling of central nervous system injuries. Neurorehabil Neural Repair 2005;19:219–26. Bilgen M. Inductively-overcoupled coil design for high resolution magnetic resonance imaging. Biomed Eng Online 2006;5:3. Bilgen M, Al-Hafez B. Comparison of spinal vasculature in mouse and rat: investigations using MR angiography. Neuroanatomy 2006;5:12–8. Bloom FE, Reilly JF, Redwine JM, Wu CC, Young WG, Morrison JH. Mouse models of human neurodegenerative disorders: requirements for medication development. Arch Neurol 2005;62:185–7. Bonny JM, Gaviria M, Donnat JP, Jean B, Privat A, Renou JP. Nuclear magnetic resonance microimaging of mouse spinal cord in vivo. Neurobiol Dis 2004;15:474–82. Farooque M, Suo Z, Arnold PM, Wulser MJ, Chou CT, Vancura RW, et al. Gender-related differences in recovery of locomotor function after spinal cord injury in mice. Spinal Cord 2006;44:182–7. Ford JC, Hackney DB, Joseph PM, Phelan M, Alsop DC, Tabor SL, et al. A method for in vivo high resolution MRI of rat spinal cord injury. Magn Reson Med 1994;31:218–23. Ghasemlou N, Kerr BJ, David S. Tissue displacement and impact force are important contributors to outcome after spinal cord contusion injury. Exp Neurol 2005;196:9–17. Guertin PA. Paraplegic mice are leading to new advances in spinal cord injury research. Spinal Cord 2005;43:459–61. Kwon BK, Oxland TR, Tetzlaff W. Animal models used in spinal cord regeneration research. Spine 2002;27:1504–10. Rosenzweig ES, McDonald JW. Rodent models for treatment of spinal cord injury: research trends and progress toward useful repair. Curr Opin Neurol 2004;17:121–31. Silver X, Ni WX, Mercer EV, Beck BL, Bossart EL, Inglis B, et al. In vivo 1H magnetic resonance imaging and spectroscopy of the rat spinal cord using an inductively-coupled chronically implanted RF coil. Magn Reson Med 2001;46:1216–22. Steward O, Schauwecker PE, Guth L, Zhang Z, Fujiki M, Inman D, et al. Genetic approaches to neurotrauma research: opportunities and potential pitfalls of murine models. Exp Neurol 1999;157:19–42. Wamil AW, Wamil BD, Hellerqvist CG. CM101-mediated recovery of walking ability in adult mice paralyzed by spinal cord injury. Proc Natl Acad Sci USA 1998;95:13188–93. Wirth 3rd ED, Mareci TH, Beck BL, Fitzsimmons JR, Reier PJ. A comparison of an inductively coupled implanted coil with optimized surface coils for in vivo NMR imaging of the spinal cord. Magn Reson Med 1993;30:626– 33.
Magnetic resonance microscopy of spinal cord injury in mouse using a miniaturized implantable RF coil Mehmet Bilgen ∗ Hoglund Brain Imaging Center and Department of Molecular and Integrative Physiology, The University of Kansas Medical School, Mail Stop 3043, 3901 Rainbow Blvd, Kansas City, KS 66160, United States Received 27 March 2006; received in revised form 26 May 2006; accepted 27 June 2006
Abstract A magnetic resonance neuroimaging method is described for high-resolution imaging of spinal cord injury in live mouse. The method is based on a specially designed radio frequency coil system formed by a combination of an implantable coil and an external volume coil. The implantable coil is a 5 mm × 10 mm rectangular design with a 9.1 pF capacitor and 22 gauge copper wire and optimal for surgical implantation over the cervical or thoracic spine. The external volume coil is a standard birdcage resonator. The coils are inductively overcoupled for imaging the spinal cord at 9.4 T magnetic field strength. The inductive overcoupling provides flexibility in tuning the resonant frequency and matching the impedance of the implanted coil remotely using the tuning and matching capabilities of the volume coil. After describing the implementation of the imaging setup, in vivo data are gathered to demonstrate the imaging performance of the coil system and the feasibility of performing MR microscopy on injured mouse spinal cord. © 2006 Elsevier B.V. All rights reserved. Keywords: Magnetic resonance imaging; Microscopy; Inductive overcoupling; Radio frequency coil; Implantable coil; Spinal cord
1. Introduction Rodent models are the backbone of neuroscience research in experimental spinal cord injury (SCI) (Kwon et al., 2002, and the references therein). According to the current trend, however, the focus is shifting from rat to paraplegic murine models because of the availability of genetically engineered transgenic and mutant mouse strains along with molecular tools (Guertin, 2005; Rosenzweig and McDonald, 2004; Steward et al., 1999). The growing number of such efforts is likely to produce new targets for therapeutic strategies aimed at promoting neuronal protection, reorganization and functional recovery following SCI in humans (Bloom et al., 2005). Magnetic resonance imaging (MRI) modalities have long been used to remotely probe the spinal cord (SC) in rodent models, and radio frequency (RF) coil systems were developed with a special feature that allowed the imaging coil to be surgically implanted at the level of the SCI in rats (Bilgen et al., 2001; Bilgen, 2004; Ford et al., 1994; Silver et al., 2001; Wirth et al., 1993). The implantable coils offer improved signal-to∗
Tel.: +1 913 588 9079; fax: +1 913 588 9071. E-mail address: [email protected].
0165-0270/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jneumeth.2006.06.024
noise ratio and small field-of-view selection due to the small dimensions of the coil and compact footprint of the RF excitation/signal detection. Both of these features aid in the acquisition of high-resolution images from the tissue of interest only. In our studies, this arrangement provides a practical solution for spatially localized imaging of the SCI and enhances the quality of the MR data acquired in longitudinal studies. Signal transmission and reception with the implanted coil is achieved via mutual-inductive coupling with a larger coil placed external to the body. We have shown that strengthening the mutual coupling increases the functionality of the implantable coil (Bilgen, 2006). Inductive overcoupling is the condition of electromagnetic interaction between two coils beyond the point at which the resonance peak of the response curve for the current in the coupled coil splits. This allows the frequency response characteristics of the implantable coil with fixed electrical elements to be modified as needed via an external coil with tuning and matching capacity. Such ability provides an environment for reliable and repeatable acquisitions of MR data, which is a highly desirable feature for studying the progression of the pathology in injured SC over time on the same animal. Previously, we have shown that an implantable coil developed for imaging rat SC can also serve as a surface coil to
94
M. Bilgen / Journal of Neuroscience Methods 159 (2007) 93–97
Fig. 1. In vivo sagittal image of a mouse spinal cord injured (arrow) at the thoracic level T11. The image was acquired using an inductively coupled surface coil developed originally as an implantable coil for imaging rat SCI. The acquisition parameters were TR /TE = 2500/12 ms, FOV=30 mm X 15 mm, image matrix = 256 X 128, slice thickness = 0.5 mm and NEX = 2. Fig. 3. Implantable coil and standard quadrature volume coil (4 cm ID) used in these studies. Note that only one channel of the volume coil is used for imaging and the other channel was terminated with 50 ohm load. The arrows point to the 9.1 pf capacitor. A dime was also included in the picture for the purpose of providing relative comparison between the sizes of the coils.
Fig. 2. Sagittal view of the mouse spinal cord at the thoracic and cervical levels. The image was acquired using a quadrature volume coil with 4 cm ID, shown in Fig. 3. The acquisition parameters were TR /TE = 4000/12 ms, FOV = 60 mm X 24 mm, image matrix = 256 X 128, slice thickness = 0.5 mm and NEX = 2.
image normal (Bilgen et al., 2005; Bilgen and Al-Hafez, 2006) and pathological SC (Bilgen, 2004) in mouse. As shown in Fig. 1, this configuration exhibits a region of high sensitivity over the thoracic levels from T10 to T12, and therefore is capable of producing excellent MR data, but only from this spatiallylimited region of the spine. However, in experimental studies, the cord is typically injured at mid-thoracic (Farooque et al., 2006; Ghasemlou et al., 2005) or cervical levels (Anderson et al., 2004). The cord at these levels can be visualized using a volume coil as shown in Fig. 2, but at lower resolution than is possible with the implantable coil and with compromised image quality. In the present study, we investigate the potential of using an implantable coil to improve the visualization of the mouse SC, as has been done in the previous rat studies. To accomplish this task, we introduce a miniaturized implantable coil for experimental mouse SCI studies at 9.4 T magnetic field strength. In the following, we describe the design and components of this coil in detail, and show its performance with data acquired from injured cords in vivo when it is inductively overcoupled to a volume coil with tuning and matching circuitry. We also discuss the benefits and limitations of this coil system in the context of tradeoffs in longitudinal studies. 2. Methods 2.1. RF coil setup We fabricated an implantable coil using a piece of circular (22 gauge, CDA 102) soft bare copper wire (MWS Wire
Industries, Westlate Village, CA) that was bent to form a rectangular “pitched rooftop” shape with dimensions 5 mm × 10 mm. A small 9.1 pF ceramic chip capacitor (American Technical Ceramics, Huntington Station, NY) was soldered at the ends of the wire to complete the circuit, as shown in Fig. 3. The coil was coated with biologically inert silicone elastomer MDX44210 (Factor II, Inc., Lakeside, AZ) for electrical isolation from its surroundings when implanted. The frequency response of the implantable coil was measured using a rectangular (25 mm × 40 mm) pickup coil and a frequency sweeper (Morris Instruments, Inc., Ottawa, Ontario, Canada). The pickup coil was tuned at about 1 GHz and therefore, did not affect the frequency response characteristics of the implantable coil. The coated coil was then sterilized by immersion in Cidex disinfectant for 24 h, followed by immersion in distilled water for 2 h before it was surgically implanted into the animal. For signal transmission/reception with the implantable coil, we employed an inductively overcoupled configuration with a standard quadrature volume coil (Fig. 3), similar to that described previously in rat studies, but the volume coil was for mouse and had a smaller inner diameter of 4 cm (Bilgen, 2006). We terminated one channel of the coil to make it operate as a linear coil. The combination of this volume coil with the implantable coil was used to image the mouse SC. 2.2. Animal surgeries for SCI and coil implantation The experiments were performed on 4-month old C57Bl/6 mice weighing between 20 and 25 g under the protocol approved by the University of Kansas Medical Center Institutional Animal Care and Use Committee. The animal was initially anesthetized with 5% isoflurane in an induction chamber, and then moved to a surgery table, where anesthesia was maintained with a mixture (1.5% isoflurane, 30% oxygen and air) delivered through a nose mask. A midline incision was made from T5 to T12 and the underlying tissue and muscle were dissected. Laminectomy was performed at the level T9. The exposed SC was moderately injured using a 0.5 mm diameter circular-tip injury bit attached to the injury device described previously (Bilgen, 2005). Other
M. Bilgen / Journal of Neuroscience Methods 159 (2007) 93–97
biomechanical parameters for inflicting the injury were set as velocity = 0.75 m/s, displacement = 0.5 mm and compression duration = 100 ms. Next, the implantable coil was centered at the injury site and held in place temporarily to assess which spinous processes needed to be removed to create enough room to stabilize the coil on the rib cage. Once the coil was placed properly on the rib cage, the overlaying paraspinal muscles were sutured at the midline to permanently hold the coil in place and the skin was closed tightly. 2.3. Magnetic resonance imaging MRI scans were performed on a 9.4 T horizontal bore scanner (Varian, Inc., Palo Alto, CA) with 31-cm bore and 400 mT/m gradient coil set. After the surgeries, the animal with the implanted coil was put in the supine position on a plastic cylinder that was cut in half along its long axis. The cylinder was inserted into the volume coil and the implanted coil was centered. To improve the coupling, the volume coil was rotated slightly with respect to the cylinder until two peaks appeared near the proton resonant frequency of 400 MHz on the sweeper’s display. The tuning and matching rods of the volume coil were then engaged to further improve the impedance matching and frequency tuning at resonance. The Plexiglas cradle, supporting the animal and coil assembly, was inserted into the scanner bore, and its placement at the magnet isocenter was confirmed with scout images. Gradient echo (GE) images were acquired in the sagittal and coronal planes to demonstrate the signal patterns associated with the implantable and volume coils over a large field-of-view (FOV). The parameters for these acquisitions were TR /TE = 40 ms/3 ms, flip angle = 45◦ , image matrix = 128 × 128, slice thickness = 2 mm and NEX = 1 and FOV = 35 mm × 50 mm, which resulted in a total acquisition time of 6 s. Next, proton density-weighted images were acquired in high-resolution, multi-slice and interleaved mode in the axial and sagittal, planes using standard spin echo (SE) sequence over smaller FOV. The acquisition parameters for the sagittal images were TR /TE = 4000 ms/12 ms, FOV = 15 mm × 10 mm, image matrix = 256 × 128, slice thickness = 0.5 mm and NEX = 2. The parameters for the axial slices were TR /TE = 4000 ms/12 ms, FOV =6 mm × 6 mm, acquisition matrix = 128 × 128, slice thickness = 1 mm and NEX = 4. The acquisition times for these scans were about 18 and 35 min, respectively. While in the scanner, the mouse was monitored using an MRcompatible small animal monitoring and gating system (Model 1025, SA Instruments, Inc., Stony Brook, NY). This system was also used for respiratory-gated acquisition to minimize the breathing-related image artifacts. The temperature of the animal was kept at 37 ◦ C by circulating warm air with 40% humidity through a 5 cm diameter plastic tube fitted at the back door of the magnet bore. 3. Results and discussion The mouse spine exhibits strong curvature, as shown in Fig. 2. While the cervical and upper thoracic sections of the spine
95
anatomically lie deep in the body, the lower thoracic spine is situated more superficial. In the adult mouse, the cervical spine together with the back of the skull measures about 11 mm in length. Except at its most upper and lower segments, the midthoracic spine remains mostly straight between the levels T5 and T11, measuring about 13 mm in length. The 10 mm length implantable coil in Fig. 3 is sized well for implantation over either the cervical or mid-thoracic segments and can be used for imaging the SC injured at these levels. But, it may not be as easy to implant the coil to image the upper thoracic spine due to the strong curvature at this level. In our experiments, the mice have tolerated the surgical procedures and the SCI at the T9 level. However, due to the small size of the mouse and the challenging environment for surgery, proper implantation of the coil over the tight spinal space required surgical skills. Before the coil was implanted, its frequency response was measured in bench tests in open air using the rectangular pickup coil. The implantable coil was observed to tune at 403 MHz. After implantation into the back of the mouse, the frequency response was measured again using the same pickup coil. This time, the tuning frequency read 398 MHz and the resonant peak was broader. It is typical to see a shift and broadening of the peak in the frequency response of the implantable coil when loaded with tissue impedance after the implantation (Bilgen, 2004). We applied thick elastomer coating on the coil to reduce the dielectric losses by increasing the physical separation between the coil and the surrounding environment. This simple approach made the coil less sensitive to changes in its surroundings and provided stability when it was implanted. When the mouse was placed into the volume coil, the resonance peak was split as expected due to the inductive overcoupling between the volume coil and implantable coil. The first peak was at 387 MHz. The second peak was observed to be closer to the resonance frequency of the scanner. Therefore, this second peak was tuned to 400 MHz by adjusting the matching and tuning circuitries of the volume coil. Fig. 4 shows the whole-body images of the mouse with the coil implanted in the sagittal and coronal views. During transmission, the currents in both the volume and implanted coils produced an excitation field in the body of the mouse. The intensity patterns in the figure depict the signals of both the footprint of the implanted coil and that of the background regions of the mouse body. The homogeneous weak signal in the background was induced by excitation from the volume coil. The inhomogeneous hyperintensity field seen near the site of the implanted coil was formed by both coils, but with a stronger contribution from the implanted coil. Acquisition with a lower FOV around the implanted coil produced the high-resolution images in Figs. 5 and 6. Laminectomized vertebrae as well as the spatial location and extent of the injury along the cord can easily be identified on the thin-sliced sagittal images in Fig. 5. The injured SC tissue is viewed in greater detail at 47 m resolution axial images in Fig. 6. These data clearly demonstrate the capability of the present approach to delineate the gray matter, surrounding white matter and cerebrospinal fluid in normal regions of the cord. More importantly, the injury is clearly depicted as patchy areas of hypointensity, representing hemorrhage in the SC tissue.
96
M. Bilgen / Journal of Neuroscience Methods 159 (2007) 93–97
Fig. 5. In vivo sagittal images of an injured mouse spinal cord acquired with the implantable coil (TR /TE = 4000/12 ms, FOV = 15 mm X 10 mm, acquisition matrix = 256X128, slice thickness = 0.5 mm and NEX = 4). Arrow points to the laminectomy and injury at T9.
Fig. 4. In vivo gradient echo images of a mouse in (a) sagittal and (b) coronal planes acquired with implantable coil that was inductively-overcoupled to the volume coil on the day of surgical implantation. The sagittal image shows the mouse body placed in supine position. The images were acquired at large FOV to demonstrate the signal patterns associated with each coil. The intensities in the images were windowed and scaled to enhance the background signal. Notice that, during transmission, the volume coil generates an excitation field that leads to the hypointense signal (long arrows) seen in the background. The implanted coil generates an excitation field that produces the hyperintense footprint (short arrows). Arrowheads point to the artifacts from heart motion. The acquisition ◦ parameters were TR /TE = 40 ms/3 ms, flip angle = 45 , image matrix = 128 X 128, slice thickness = 2 mm and NEX = 1 and FOV = 35 mm X 50 mm.
Although the inductively overcoupled coil system is capable of providing high quality images, it also poses a limitation because of the physics involved in the MR data acquisition over small FOV (Bilgen, 2006). That is the images acquired inherently contain wraparound signals from tissue outside the selected FOV that has been excited by the volume coil. As shown in Fig. 4, strong signal artifacts caused by motion of the heart are present on the GE images, but signals from other motionless regions are significantly low in intensity. Compared to the GE
data, the SE acquisition at a large FOV produces lower background signal. The consequence of the wraparound is that separating the signals on the regions of interest from the background noise becomes difficult, even after the contrast and brightness of the displayed image are adjusted. Since respiratory gating was employed during our acquisitions, the wraparound artifacts contributing to the SE images in Figs. 5 and 6 were only those from the heart motion, and minimal because of their low signal intensities. In Fig. 7, we provide images of an injured SC acquired ex vivo (i.e., the spine was dissected but the vertebral body was left intact) using the implantable coil that was inductively coupled to the rectangular pickup coil, instead of the volume coil. These images establish the upper boundary on the image quality under the ideal condition of artifact-free acquisition; thereby demonstrating what can ultimately be achievable in terms of imaging the mouse SC. It is interesting to note that the data in Figs. 5 and 6 are comparable to these artifact-free benchmark images in quality, clearly demonstrating the capability of our approach to accomplish the high-resolution imaging task of the SCI in mouse. The approach yields images of adequate quality, spatial sensitivity and specificity, while maintaining good intensity contrast, which makes it easy to differentiate between the injury and the other structures within the SC. Our group as well as others is increasingly using mouse models in experimental SCI studies. However, in vivo MRI of the
Fig. 6. In vivo axial images of the injured mouse in Fig. 5 (TR /TE = 4000/12 ms, FOV = 6 mm X 6 mm, acquisition matrix = 128X128, slice thickness = 1 mm and NEX = 4).
M. Bilgen / Journal of Neuroscience Methods 159 (2007) 93–97
97
Fig. 7. Ex vivo axial images of an injured mouse spinal cord acquired with the implantable coil in Fig. 3 while the vertebral body was still intact (TR /TE = 4000/12 ms, FOV = 4 mm X 6 mm, acquisition matrix = 128X128, slice thickness =0.5 mm, slice gap = 0.5 mm and NEX = 4).
injured mouse SC has received little attention to date (Bilgen et al., 2005; Bilgen and Al-Hafez, 2006; Bonny et al., 2004; Wamil et al., 1998). The previous studies imaged the mouse SC with limited quality and spatial sensitivity by using surface or volume coils. In this paper, we addressed this issue and described an inductively overcoupled coil system consisting of a miniaturized implantable coil. Fabricating this coil was simple and low-cost, and cutting and forming the copper wire and soldering the inexpensive ceramic chip capacitor took less than 15 min. The volume coil was a standard quadrature coil available in any MRI laboratory. The tuning and matching feature of the coil system made it flexible and practical to readjust any changes in the frequency response after its implantation. The results presented above clearly demonstrate the feasibility of acquiring high quality in vivo MR data from the injured SC in mice at 9.4 T using this coil system. Although, the data presented above were anatomical visualization, future studies may take advantage of this coil in using the structural and functional MR modalities, such as diffusion tensor and contrast enhanced imaging modalities, to characterize the longitudinal changes in the pathobiology of the SCI in mice. Acknowledgments The author thanks Dr. Baraa Al-Hafez and Dr. Yong-Yue He for their help in animal preparation, surgeries for coil implantation and MRI scans. References Anderson KD, Abdul M, Steward O. Quantitative assessment of deficits and recovery of forelimb motor function after cervical spinal cord injury in mice. Exp Neurol 2004;190:184–91. Bilgen M, Elshafiey I, Narayana PA. In vivo magnetic resonance microscopy of rat spinal cord at 7 T using implantable RF coils. Magn Reson Med 2001;46:1250–3. Bilgen M. Simple, low-cost multipurpose RF coil for MR microscopy at 9.4 T. Magn Reson Med 2004;52:937–40.
Bilgen M, Al-Hafez B, Berman NE, Festoff BW. Magnetic resonance imaging of mouse spinal cord. Magn Reson Med 2005;54:1226–31. Bilgen M. A new device for experimental modeling of central nervous system injuries. Neurorehabil Neural Repair 2005;19:219–26. Bilgen M. Inductively-overcoupled coil design for high resolution magnetic resonance imaging. Biomed Eng Online 2006;5:3. Bilgen M, Al-Hafez B. Comparison of spinal vasculature in mouse and rat: investigations using MR angiography. Neuroanatomy 2006;5:12–8. Bloom FE, Reilly JF, Redwine JM, Wu CC, Young WG, Morrison JH. Mouse models of human neurodegenerative disorders: requirements for medication development. Arch Neurol 2005;62:185–7. Bonny JM, Gaviria M, Donnat JP, Jean B, Privat A, Renou JP. Nuclear magnetic resonance microimaging of mouse spinal cord in vivo. Neurobiol Dis 2004;15:474–82. Farooque M, Suo Z, Arnold PM, Wulser MJ, Chou CT, Vancura RW, et al. Gender-related differences in recovery of locomotor function after spinal cord injury in mice. Spinal Cord 2006;44:182–7. Ford JC, Hackney DB, Joseph PM, Phelan M, Alsop DC, Tabor SL, et al. A method for in vivo high resolution MRI of rat spinal cord injury. Magn Reson Med 1994;31:218–23. Ghasemlou N, Kerr BJ, David S. Tissue displacement and impact force are important contributors to outcome after spinal cord contusion injury. Exp Neurol 2005;196:9–17. Guertin PA. Paraplegic mice are leading to new advances in spinal cord injury research. Spinal Cord 2005;43:459–61. Kwon BK, Oxland TR, Tetzlaff W. Animal models used in spinal cord regeneration research. Spine 2002;27:1504–10. Rosenzweig ES, McDonald JW. Rodent models for treatment of spinal cord injury: research trends and progress toward useful repair. Curr Opin Neurol 2004;17:121–31. Silver X, Ni WX, Mercer EV, Beck BL, Bossart EL, Inglis B, et al. In vivo 1H magnetic resonance imaging and spectroscopy of the rat spinal cord using an inductively-coupled chronically implanted RF coil. Magn Reson Med 2001;46:1216–22. Steward O, Schauwecker PE, Guth L, Zhang Z, Fujiki M, Inman D, et al. Genetic approaches to neurotrauma research: opportunities and potential pitfalls of murine models. Exp Neurol 1999;157:19–42. Wamil AW, Wamil BD, Hellerqvist CG. CM101-mediated recovery of walking ability in adult mice paralyzed by spinal cord injury. Proc Natl Acad Sci USA 1998;95:13188–93. Wirth 3rd ED, Mareci TH, Beck BL, Fitzsimmons JR, Reier PJ. A comparison of an inductively coupled implanted coil with optimized surface coils for in vivo NMR imaging of the spinal cord. Magn Reson Med 1993;30:626– 33.