Three-dimensional visualization of protein expression in mouse brain structures using imaging mass spectrometry

Anna C. Crecelius 0 1 2 D. Shannon Cornett 0 1 2 Richard M. Caprioli 0 1 2 Betsy Williams 0 1 2 Benoit M. Dawant 0 1 2 Bobby Bodenheimer 0 1 2 0 Published online May 31, 2005 Address reprint requests to Dr. R. M. Caprioli, Mass Spectrometry Research Center, Vanderbilt University , 465 21st Avenue S, Room 9160, Medical Research Building III , Nashville, TN 37232, USA 1 Department of Electrical Engineering and Computer Science, Vanderbilt University , Nashville, Tennessee, USA 2 Mass Spectrometry Research Center and Department of Biochemistry, Vanderbilt University , Nashville, Tennessee, USA We have developed a method to visualize matrix-assisted laser desorption ionization imaging mass spectrometry (MALDI IMS) data aligned with optically determinable tissue structures in three dimensions. Details of the methodology are exemplified using the 3-D reconstruction of myelin basic protein (MBP) in the corpus callosum of a mouse brain. In this procedure, optical images obtained from serial coronal sections are first aligned to each other to reconstruct a surface of the corpus callosum from segmented contours of the aligned images. The MALDI IMS data are then coregistered to the optical images and superimposed into the surface to create the final 3-D visualization. Correlating proteomic data with anatomical structures provides a more comprehensive understanding of healthy and pathological brain functions, and holds promise to be utilized in more complex anatomical arrangements. (J Am Soc Mass Spectrom 2005, 16, 1093-1099) 2005 American Society for Mass Spectrometry - W(CT, positron emission tomography (PET) ith the introduction of computer tomography and magnetic resonance imaging (MRI, 3-D medical imaging has become important in studying anatomical, physiological, and functional information [1]. However, these in vivo tomographic imaging techniques do not have the ability to display protein distributions. Histochemical techniques provide 2-D maps of protein distributions, but require an antibody for visualizing each known protein, and only a limited number of proteins can be visualized in each slice using this technique. In recent years, a number of mass spectrometry techniques have been used to obtain the x/y spatial localization of many compounds on sample surfaces [2 6]. Of the current techniques, MALDI IMS provides the widest m/z range of species which can be imaged and is most commonly used to image protein distributions in thin sections of tissue. Using the technique, imaging is accomplished by acquiring a mass spectrum at discrete locations along a grid pattern prescribed over the surface. The resulting data set includes the x, y coordinates (pixel) and the corresponding spectra. By extracting the measured intensity values for a particular m/z plane from the data set and plotting on a color scale, 2-D ion images can be created. Expanding this technique to include images of serial sections from a single specimen provides a depth dimension to the data set, allowing the 3-D representation of even unidentified proteins in their full spatial and multi-dimensional distribution. Different image modalities are often combined in neuroimaging to assist in the understanding of brain functions in normal and diseased stages [79]. Unlike in vivo tomographic imaging techniques, 2-D modalities such as optical, histological, and MALDI IMS images require the specimen to be sliced into thin sections, which can produce tissue tearing and deformation. Therefore, correlating consecutive 2-D images to obtain a 3-D reconstruction is more demanding. In order to reconstruct a 3-D visualization involving data from different modalities, the establishment of a correspondence between the modalities through image registration techniques [10] is required. The present study focuses on developing the 3-D visualization of proteomic data correlated to anatomical features in the brain. In the present example, we chose to model the distribution of myelin basic protein (MBP) in the corpus callosum (CC). The mouse brain was chosen for our first model because of its small size and because an atlas exists which can be used as reference for the methodology development. To create the 3-D model of MBP within the CC, optical, histological, and MALDI IMS images were acquired and processed according to the workflow illustrated in Figure 1. A number of registration steps were required to produce the final 3-D protein distribution. First, optical micrographs of unstained sections were aligned to a reference atlas. Although not explicitly necessary, use of the external reference improves the consistency of our modeling and eliminates the step of generating our own reference. Next, the optical images and MALDI IMS images were registered to each other based on landmarks visible in both imaging modalities. Finally, a 3-D surface of the anatomical feature was created from the aligned optical images and the created 2-D MALDI IMS images were inserted into the 3-D surface model according to their z coordinate position. The resulting 3-D model provides, for the first time, a unique correlated view of MBP distribution within the CC of mouse brain. However, the methodology can be used to map the distribution of any protein visible in the ion image spatially correlated with anatomical features observed in optical micrographs. Specimen Preparation and Cryosectioning The brain of a male, 6-week old C57Bl/6J mouse was removed from the skull, loosely wrapped in aluminum foil, and immediately frozen by slow immersion into liquid nitrogen for several seconds. The brain was stored at 80 C until sectioning in order to minimize protein degradation caused by temperature and oxidation[11].Theunfixedbrainwassectionedat 16C into successive coronal sections spanning the corpus callosum using a cryostat (CM3050S Leica, Nussloch, Germany). Each section was thaw-mounted onto indium-tin oxide coated glass slides (Delta Technologies Ltd., Stillwaters, MN) and stored at 80 C until time of analysis. Optical Image Acquisition Prior to MALDI IMS analysis, the glass slide with the desired mouse brain section was thawed in a vacuum desiccator for 15 min to avoid moisture condensation that could cause delocalization of proteins. Four black ink dots (diameter: 800 m, printed on an adhesive label using a laser printer, were attached to the slide framing the brain slice for use as registration landmarks. An optical image of the section and surrounding landmarks was acquired using a 640 480 digital capture board coupled to an analog camera. The image resolution was 35 m/pixel and was stored to disk in tiff format. Coating Brain Section with MALDI Matrix Paper masking was applied over the ink spots, and the brain section was coated with MALDI matrix. The matrix (sinapinic acid, 20 mg/ml in acetonitrile/water/ TFA 50/50/0.3, vol/vol) was deposited uniformly over the entire mouse brain section using a TLC reagent sprayer (Fisher Scientific, Suwanee, (...truncated)