Magnetic Resonance Microscopy (MRM) of Single Mammalian Myofibers and Myonuclei

Recently, the first magnetic resonance microscopy (MRM) images at the cellular level in isolated mammalian brain tissues were obtained using microsurface coils. These methods can elucidate the cellular origins of MR signals and describe how these signals change over the course of disease progression and therapy. In this work, we explore the capability of these microimaging techniques to visualize mouse muscle fibers and their nuclei. Isolated myofibers expressing lacZ were imaged with and without a stain for β-galactosidase activity (S-Gal + ferric ammonium citrate) that produces both optical and MR contrast. We found that MRM can be used to image single myofibers with 6-μm resolution. The ability to image single myofibers will serve as a valuable tool to study MR properties attributed to healthy and myopathic cells. The ability to image nuclei tagged with MR/Optical gene markers may also find wide use in cell lineage MRI studies.

The ability to monitor the onset and progression of disease, as well as evaluate therapeutic efficacy at the cellular level in a non-invasive and non-destructive manner, contributes to the understanding of disease etiology and provides more information for clinicians. To date, the most prevailing methodology to assess cellular status is through fluorescent microscopy techniques [1][2][3][4] . Through the use of reporter genes cells have been imaged at the whole body and tissue level using positron emission tomography, fluorescent mediated tomography, and luminescence where traditional marker genes and fluorescent protein are commonly implemented for detection on the cellular level. However, light-based technologies possess inherent underlying limitations, primarily due to photon attenuation in the visible wavelength range, limiting penetration depth within biological samples.
As a complementary methodology, magnetic resonance imaging (MRI), a versatile, non-ionizing, and non-invasive diagnostic imaging modality, has been employed to study the distribution of water molecules in the muscle cells of barnacles 5,6 , frogs 7,8 , rats [9][10][11] , and humans 12 . Due to recent advances in hardware and software which have improved spatial resolution in MRI, magnetic resonance microscopy (MRM) has transitioned from a technique used to image single, large cells, such as frog eggs 13 and neurons of gastropod mollusks [14][15][16][17][18] , to much smaller cells such as mammalian α -motor neurons in rats 19 , pigs and humans 20 . Furthermore, it has recently proven its capability of visualizing the 3D brain connectivity in the fly brain based on endogenous contrast, which is the first map from the whole animal head 21 . By utilizing MR contrast mechanisms such as NMR relaxation and diffusion, MRM can take advantage of endogenous subcellular contrast, complementing other microscopy techniques.
In vivo cellular tracking is achieved through a combination of novel exogenous markers and state-of-the-art imaging techniques. Specifically, iron and gadolinium based contrast agents have been utilized to enhance spin-spin (T 2 ) and spin-lattice (T 1 ) relaxivity, which contain iron and gadolinium based products, respectively. Cellular iron-based contrast agents have been used to monitor gene regulation by upregulating and/or overexpressing proteins that bind cellular iron and result in changes in the MR signal through locally induced magnetic Scientific RepoRts | 7:39496 | DOI: 10.1038/srep39496 field inhomogeneities. These inhomogeneities in turn lead to an amplification of negative contrast and detectability by the reduction of T 2 or T 2* relaxation times 22,23 . This approach can also be used to track cells exogenously labeled with superparamagnetic particles 24,25 . Due to the increased induced magnetic field inhomogeneities this has been used to detect single cells in vitro 26 and in vivo 27 . It can also be used to assess transgene expression using the beta-galactosidase reporter system in translational research studies 22,28-30 as well as cancer clinical trials 31,32 . Other commonly employed cellular contrast agents include gadolinium-loaded nanoparticles, which increase T 1 relaxation rates of neighboring water molecules 33,34 . However, both contrast agents possess shortcomings in demarcating spatial characteristics at the cellular level due to the low specificity and spatial resolution of MRI. Such limitations warrant further investigation into these and alternative contrast agents including confirmation of their ability to accurately represent spatial relationships in the microscopic domain by correlative histological analysis 35 .
In this study, we employed the lacZ gene reporter system under control of the mouse myosin light chain 3 F promoter/enhancer element 36 to detect β -galactosidase (β -gal) activity in cell nuclei 22,37 . Intact, single extensor digitorum longus (EDL) myofibers were harvested from wild-type control (C57/BL6) mice or transgenic mice expressing muscle-specific nuclear lacZ. Isolated fibers were labeled with S-Gal (3,4-Cyclohexenoesculetin β -D-galactopyranoside; Sigma-Aldrich, St. Louis, MO) and Ferric Ammonium Citrate (FAC). S-Gal is a commercially available histological stain for lacZ, while FAC allows for in vitro detection of β -gal myonuclei through negative contrast on T 2* -weighted MRI scans 38,39 . The stained fibers from our transgenic strain were compared with unlabeled control fibers with S-Gal and FAC. To optimize our findings, fibers were doped with different concentrations of S-Gal and FAC to determine ideal staining conditions for the myofiber nuclei. Resolution and signal-to-noise ratio (SNR) were adequate for delineation of myonuclei demonstrating the complementary role of MRM to other methods of optical microscopy. Such high signal and resolution characteristics were achieved by using state-of-the-art radiofrequency (RF) micro coils 40 in conjunction with strong and fast switching gradient coils 41 .
A representative schema of the mechanism of the experimental model is presented in Fig. 1. Because of the mouse myosin light chain (MLC) 3F promotor/enhancer element, expression of nLacZ (and subsequent translation of β -galactosidase (β -gal)) is specific to muscle. Briefly, S-Gal and FAC form Fe 3+ in the presence of β -galactosidase (β -gal) resulting in visible black precipitate 38 . Depending on the length of incubation and concentrations of substrates, optical and MR contrast can be optimized to enhance contrast.

Results
Using MRM, we were able to image single muscle fibers directly employing only native tissue contrast ( Fig. 2) with a 3D fast low-angled shot (FLASH) sequence. In this experiment, an isotropic resolution of 6 μ m and a SNR of 20 were achieved. Alternatively, using a 3D spin echo sequence, the spatial resolution was 8 × 8 × 31 μ m and the SNR was 14. MRM using 3D FLASH and spin echo sequences demonstrates native contrast generated by the muscle fibers at 6 μ m isotropic resolution (Fig. 2a) and 8 μ m in-plane resolution (Fig. 2b). A direct comparison between morphological images of isolated muscle fibers using MRM ( Fig. 2a and b) and differential interference contrast (DIC) microscopy (Fig. 2c) demonstrated the normal birefringence due to sarcomeric striations consisting of dark A-band and bright I-band in an interdigitated morphology (inset of Fig. 2c), and indicated that the fiber structures were intact. The relative scale of the muscle fiber and MR RF micro surface coil are shown in Fig. 2(d), as observed by light microscopy.
In order to confirm the identity of anatomical structures in our MR images, a genetic label capable of both histological and MRI contrast was used. Incubation with S-Gal/FAC allowed us to visualize muscle cells that were expressing lacZ by forming an opaque precipitate in the cytoplasm or nucleus. This ferric iron (Fe 3+ ) precipitate was visible in both 3D FLASH microimages ( Fig. 3a and b) and light microscopy ( Fig. 3c and d). Accumulated iron content along the individual fibers contrasted with neighboring regions due to heterogeneous uptake. Without uniformly controlled targeting or binding of the iron-containing S-Gal, the resultant T 2 (or T 2* ) contrast has been amplified mostly around the punctate areas along the fiber wall ( Fig. 3a and b). Four myofibers were selected under the light microscope for possessing maximal differential phase response before being embedded into an agarose gel for MR imaging (Fig. 3c).
MRM of single myofibers using 3D FLASH enabled visualization of myonuclei at the shortest echo time used (TE = 1.5 ms) and detection of individual nuclei with 6 μ m isotropic resolution in consecutive slices with a total thickness of 18 μ m (Fig. 4a-c, red arrowheads). The spatial distribution of nuclei was visualized using 3D segmentation (Fig. 4d). We could detect precipitate in the FAC/S-Gal-doped myofiber after only 15 minutes of staining A group of lacZ-expressing myonuclei in myofibers that were stained for 15 minutes with FAC/S-Gal was visualized using bright-field microscopy ( Fig. 5a). By employing the bigger diameter coil, i.e. 500 μ m, for the wider coverage, stained nuclei demonstrated hypointense contrast in T 2 (or T 2* ) images along the individual fibers in 3D FLASH MRM with 8 μ m resolution (Fig. 5b). The through-plane arrangement of the coil was differentiated in such a way that the closer portions of the fiber was tilted vertically in a diagonal direction to the coil surface along with the remaining fibers including the clear demarcation of the fiber placed horizontally in the middle of the field of view (FOV). Spin echo images (in-plane resolution = 8 μ m) also detected iron labeling in the projected image of entire geometry and arrangement of the myofibers, i.e., X-shaped geometrical configuration of intersection of myofibers at the center of the FOV for the landmark, but the spatial localization of individual nuclei was not possible because of the thicker through-plane resolution employed (160 μ m) ( Fig. 5c and d). By increasing the echo time from 5.8 (Fig. 5c) to 13.5 ms (Fig. 5d), the hypointense MRM signals emanating from S-Gal labeled myofibers resulted in noticeably amplified contrast. Temporal and regional changes in iron-deposition on single myofibers were compared in Fig. 6. Figure 6(a) presents the myofiber after doping with S-Gal/FAC for 5 minutes. Expectedly, the amount of S-Gal/FAC deposition on the myonuclei was comparatively less than that of the myofibers which were incubated for 15 minutes (Fig. 5a). After an incubation time of 45 minutes, the myofiber became saturated with iron-containing S-Gal/FAC resulting in obfuscation of the labeled nuclei (Fig. 6b). The myofiber became entirely iron-saturated and opaque after an incubation time of 60 minutes (Fig. 6c).

Discussion
To the best of our knowledge, this study is the first to report visualization of individual mammalian muscle fibers and nuclei using MRM methods. Muscle nuclei were highlighted by quickly decaying T 2 signals using lacZ as an anatomically targeted gene reporter following incubation with S-Gal/FAC 38,39 (Fig. 1).
To test the specificity of lacZ as a genetic reporter and the role of β -galactosidase activity, we incubated C57/ BL6 myofibers with S-Gal/FAC (Fig. 4). In the negative control fibers containing no S-Gal/FAC staining, there was no contrast in 3D FLASH (Fig. 2a) or 3D spin echo MRM (Fig. 2b). By contrast, in the stained fibers, the areas where S-Gal was taken up demonstrated T 2 (or T 2* ) contrast enhancement, primarily in punctate areas surrounding the fiber (Fig. 3a and b). However, because of the low efficiency of myonuclei targeting as well as the non-uniform distribution of iron-containing S-Gal, these techniques alone could not demarcate the precise locations of myonuclei. The occurrence of excessive background staining of iron is in strong agreement with our previous study 39 . Building upon the reported detectability of T 2 (or T 2* ) signal after S-Gal/FAC incubation of lacZ-expressing myofibers, this study is the first to address the optimization of incubation duration (15 minutes) and spatially appropriate resolution for visualizing myonuclei with MRM (6 μ m) (Fig. 4).
Subcellular resolution was achieved using dedicated gradient coils with 66 T/m strength and 1.1 T/m/A slew rate 41 . These hardware characteristics allowed us to overcome bandwidth and diffusion-based resolution limitations, as discussed previously 19,41,42 . Because of the linear relationship between iron concentration and the external magnetic field, the required minimum concentration of ferritin and exposure time to generate T 2 (or T 2* )-based contrast could be reduced 10-fold. Such concentrations of iron are smaller than those previously demonstrated 43 . This may be an additional and relevant benefit to use T 2 -based contrast agents in high field MRM rather than T 1 -based agents, as T 1 -based agents' contrast becomes smaller in proportion to the strength of the external magnetic field 44 .
Investigation into the microstructural underpinnings of subcellular compartments using endogenous MR contrast has been reported in nervous tissue systems including isolated aplysia neurons (300-500 μ m) [14][15][16][17][18]45,46 , rat 19 and human α -motor neurons 20 , and xenopus gamete cells 13,[47][48][49][50][51] . The subcellular regions of the Aplysia L7 neuron visualized using native MR contrast were recently identified through correlative histological analysis 18 . By contrast, there have not been similar studies to investigate MR signal characteristics at the subcellular level in skeletal muscle in large part due to the low sensitivity and specificity in MRI. In this study, by highlighting the efficient contrasting effect by enlarged surface area of exogenous contrast agents in isolated individual muscle cells, the microstructural origin of contrast enhanced subcellular structures such as myonuclei was identified clear enough to differentiate the linear relationship between exposure time to FAC, the empirical threshold temporal information, i.e., 15 min exposure time, and contrast enhancement for the identification of the subcellular targeting for potential therapeutic applications with fine tunability. With the technical improvement of sensitivity and resolving power of MRM, these results warrant further investigation on diagnostic utility of (sub) cellular profiles and possible role in cellular tracking based on endogenous contrast. With the evolving technical improvements to sensitivity and resolving power offered by MRM, these results suggest further investigation into the potential diagnostic utility of sub-cellular, tissue-specific labeling methods is warranted. The lacZ gene reporter system presented in this study may further expand the ability of MRM to non-invasively and longitudinally monitor transgene expression. Specifically, this technology may be useful for pre-labeling of rapidly proliferating stem cells in vitro, in order to follow their migration to their final destination in vivo. Furthermore, with natural reporter systems, direct in vivo mapping may be possible based on gene reporter expression in transgenic animal models. Such tools could lead to a better understanding of the cellular mechanisms responsible for a multitude of human pathologies including muscular dystrophies and cancers.
Regarding translational issues, we do not anticipate these spatial resolutions to be possible in the foreseeable future in human studies due to the intrinsic hardware requirements. Rather, these works will enable us to understand the cellular origins of the signal changes seen in clinical MRI (or animal studies) using these gene markers. In this study, we chose to demonstrate the use of a generic gene marker under the control of a very specific gene muscle promoter. In our experience the extrapolation to in vivo studies is not limited by the promoter but by the substrate availability of S-gal and the iron. Whereas we have shown in the past this reporter system can be used in mice 39 and others have demonstrated its use in a in vivo tumor model it relied on highly tissue permeability in vivo 30 . Human represents further challenges due to the need to introduce a marker gene without clinical benefit, the off-label use of iron supplementation, and the approval to use S-gal in humans. For these reasons, this MR reporter is best utilized as demonstrated in this study as a MR histological stain.
Conclusions and Future Work. In summary, this study has demonstrated an in vitro method for using MRM as a tool to study myonuclei of individual muscle fibers. Similar methodology may offer novel means to non-invasively track cells in vivo. This type of highly efficient cellular labeling may allow for MR-based cell tracking to serving a vital role in the study and monitoring of therapeutic interventions for a wide variety of human diseases. As an alternative to invasive biopsy, high resolution MRM is a non-destructive, non-ionizing technology that can visualize tissue dynamics using unique contrast. Such methods could provide a more useful, tissue-specific protocol for cellular tracking and potentially aid in the early detection of pathology using molecular imaging techniques. In conclusion, the findings from this study demonstrate that cellular MRM in mammalian muscle fibers is feasible, and that-when combined with histological labeling techniques-MRM allows for detection of subcellular components in mammalian tissue: myonuclei. Future studies will focus on development of protocols from molecular biology into MRI techniques for diagnostic clinical medicine.   Methods Muscle Fiber Preparation. All animal procedures were conducted in accordance with guidelines in the National Academies of Sciences' Guide for the Care and Use of Laboratory Animals and were approved by the University of Washington IACUC. Single myofibers were isolated from dissected extensor digitorum longus (EDL) muscles of wild type (C57/BL6) and transgenic MLC3FnLacZ mice. Intact EDL muscles were digested with DMEM containing 4 mg/mL Collagenase type 1 (Sigma-Aldrich, St. Louis, MO, USA) for 1 hour at 37 °C, followed by manual trituration using fire polished glass Pasteur pipettes to release individual fibers (as described in Keire et al.) 52 . Isolated fibers were fixed for 5 minutes in 2% formaldehyde and stored in phosphate-buffered saline (PBS) at 4 °C until staining and MRM. Myofibers were individually separated in a petri dish of culture medium. Then, a 1:1 molar ratio of S-Gal: FAC solution was mixed with 1% PBS and added to the culture dish containing fibers. These were allowed to incubate in a 37 °C water bath for up to 1 hour.
Sample Positioning, MRM, and 3D Visualization. All MRM was carried out on a 600 MHz (14.1 T) vertical-bore magnet (Oxford Instruments) interfaced to a Bruker Biospin console. Imaging gradient strengths up to 66 T/m were provided by a newly designed and fast-switching (1.1 T/m/A) planar gradient system (Bruker Biospin, Z110828, B6406). Initially, myofibers were embedded in low-melting point agarose Fisher) and positioned by hand with the aid of a dissecting scope (Zeiss, OPMI 1-FC). Next, an agarose block containing embedded samples was placed directly on the RF microcoil. Lastly, the tissue well was sealed using PCR film (ABgene, AB-0558) to prevent leakage and ensure the physical stability of samples. Imaging protocols were repeated for each stain concentration and incubation time tested (n = 3).
In MRM experiments targeting myonuclei, micro surface-coils of 200 μ m (inner diameter) (Bruker Biospin, B6371) and 500 μ m (inner diameter) (Bruker Biospin, B6370) were utilized. Three-dimensional FLASH datasets containing the nucleus and neighboring intracellular regions of single muscle fibers were segmented, analyzed, and reconstructed using 3D image analysis software (Amira 5.4.0; Visage Imaging) in order to visualize subcellular structures from multiple angles.