Matrix Solution Fixation: Histology-Compatible ... - ACS Publications

We have developed a fast, inexpensive, histology-compatible tissue preparation method for matrix-assisted laser desorption/ionization (MALDI)-MSI, whi...
7 downloads 0 Views 841KB Size
Download PDF
Anal. Chem. 2007, 79, 7416-7423

Matrix Solution Fixation: Histology-Compatible Tissue Preparation for MALDI Mass Spectrometry Imaging Nathalie Y. R. Agar,*,† Hong Wei Yang,† Rona S. Carroll,† Peter M. Black,† and Jeffrey N. Agar*,‡

Department of Neurosurgery, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts 02115, and Chemistry Department and Volen Center, Brandeis University, Waltham, Massachusetts 02454

Mass spectrometry imaging (MSI) acquires a grid of spatially resolved mass spectra and provides a molecular landscape of a tissue. This can have a myriad of uses: from basic tissue characterization to a comprehensive pathological diagnosis. We have developed a fast, inexpensive, histology-compatible tissue preparation method for matrix-assisted laser desorption/ionization (MALDI)MSI, which overcomes current sample preparationimposed limitations in image resolution. Tissue sections are prepared via simultaneous fixation and matrix deposition. This is accomplished by incorporating the MALDI matrix into solvents that preserve tissue integrity when applied according to standard histology procedures. This concept was expanded to include multiple histology protocols, thereby enabling analysis to be tailored to a variety of biomolecules and tissue types. Disease diagnosis and therefore treatment rely upon observations of tissue and cellular characteristics such as proliferation, cellular and nuclear morphology, vascularization, and specific biomarkers. For example, both the choice of treatment and the risks taken during surgical resection of a tumor mass in the brain are dictated by tumor type and malignancy grade. Tangentially, detailed pathology characterization is revealing considerable overlap in the molecules involved in one disease and the next, consistent with disease continuum rather than discrete disease states. For instance, the lines between Alzheimer’s, Pick’s, and Parkinson’s diseases are often blurred,1 as are the lines between gradations of a given tumor type.2-4 As the list of overlapping molecules involved in cancers and other diseases grows, so does the need for comprehensive molecular profiling as part of diagnosis.3,5 * To whom correspondence should be addressed. E-mail: N.Y.R.A. ([email protected]) or J.N.A. ([email protected]). † Brigham and Women’s Hospital. ‡ Brandeis University. (1) Galpern, W. R.; Lang, A. E. Ann. Neurol. 2006, 59, 449-458. (2) Caprioli, R. M. Cancer Res. 2005, 65, 10642-10645. (3) Schwartz, S. A.; Weil, R. J.; Thompson, R. C.; Shyr, Y.; Moore, J. H.; Toms, S. A.; Johnson, M. D.; Caprioli, R. M. Cancer Res. 2005, 65, 7674-7681. (4) Iwadate, Y.; Sakaida, T.; Hiwasa, T.; Nagai, Y.; Ishikura, H.; Takiguchi, M.; Yamaura, A. Cancer Res. 2004, 64, 2496-2501. (5) Schwartz, S. A.; Weil, R. J.; Johnson, M. D.; Toms, S. A.; Caprioli, R. M. Clin. Cancer Res. 2004, 10, 981-987.

7416 Analytical Chemistry, Vol. 79, No. 19, October 1, 2007

The recent adaptation of mass spectrometers and their respective computer applications to accommodate tissue analysis6 provides such a comprehensive molecular profile3,4,7-9 and saves time compared to histology. Mass spectrometry (MS) is a wellestablished analytical technique used to identify and characterize molecules based upon their molecular weight, and MALDI10 is the desorption/ionization method of choice for the MS analysis of large11 and/or labile biomolecules in complex biological samples. In the MALDI-MSI experiment, hundreds of closely spaced MALDI-MS spectra are taken in a grid pattern where each spectrum is analogous to a pixel.12-14 Each pixel contains information on the mass and intensity of hundreds of biomolecules, which can be translated into a spatial map of molecular distribution and abundance. MALDI requires a photon-absorbing “matrix” compound to enhance desorption.10,15 To prepare samples for MSI, the matrix is dissolved and this solution is applied to tissues. The typical solvent used to dissolve the matrix, 50% acetonitrile/50% aqueous 0.1% trifluoroacetic acid (50% ACN/50% aq 0.1% TFA), also solubilizes proteins, such that the application of matrix solution to tissue is thought to delocalize proteins12 and disturb tissue integrity.16 Delocalization of proteins has a deleterious effect on image quality, and thus strategies for minimizing protein delocalization have been developed. Devices that can deposit consistent amounts of matrix solution in discrete locales, such as automated acoustic deposition, afford spatial resolution of ap(6) Stoeckli, M.; Chaurand, P.; Hallahan, D. E.; Caprioli, R. M. Nat. Med. 2001, 7, 493-496. (7) Rahman, S. M.; Shyr, Y.; Yildiz, P. B.; Gonzalez, A. L.; Li, H.; Zhang, X.; Chaurand, P.; Yanagisawa, K.; Slovis, B. S.; Miller, R. F.; Ninan, M.; Miller, Y. E.; Franklin, W. A.; Caprioli, R. M.; Carbone, D. P.; Massion, P. P. Am. J. Respir. Crit. Care Med. 2005. (8) Yanagisawa, K.; Shyr, Y.; Xu, B. J.; Massion, P. P.; Larsen, P. H.; White, B. C.; Roberts, J. R.; Edgerton, M.; Gonzalez, A.; Nadaf, S.; Moore, J. H.; Caprioli, R. M.; Carbone, D. P. Lancet 2003, 362, 433-439. (9) Chaurand, P.; Cornett, D. S.; Caprioli, R. M. Curr. Opin. Biotechnol. 2006, 17, 431-436. (10) Karas, M.; Bachmann, D.; Bahr, U.; Hillenkamp, F. Int. J. Mass. Spectrom. Ion Processes 1987, 78, 53-68. (11) Spengler, B.; Kaufmann, R. Analusis 1992, 20, 91-101. (12) Gusev, A. I.; Vasseur, O. J.; Proctor, A.; Sharkey, A. G.; Hercules, D. M. Anal. Chem. 1995, 67, 4565-4570. (13) Stoeckli, M.; Farmer, T. B.; Caprioli, R. M. J. Am. Soc. Mass Spectrom. 1999, 10, 67-71. (14) Caprioli, R. M.; Farmer, T. B.; Gile, J. Anal. Chem. 1997, 69, 4751-4760. (15) Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2299-2301. (16) Aerni, H. R.; Cornett, D. S.; Caprioli, R. M. Anal. Chem. 2006, 78, 827834. 10.1021/ac071460e CCC: $37.00

© 2007 American Chemical Society Published on Web 09/07/2007

proximately 250 µm.16 A recent alternative to droplet deposition relies upon a tissue-washing procedure in organic solvents, followed by the immersion of the entire tissue in a typical matrix dissolution solvent (50% ACN/50% aq 0.1% TFA).17 However, since proteins are known to be soluble in such solutions, their ability to limit protein diffusion and maintain spatial resolution have been questioned.16 We initially prepared tissue sections for MALDI-time-of-flight (TOF) MSI using both manual spotting and modified electrospraying approaches.18 Although each method can produce images, image quality was inconsistent and spatial resolution was operator dependent. In an effort to improve upon frozen tissue sample preparation time and quality for imaging, we developed a method that we refer to as the matrix solution fixation (MSF) method. MSF is a histology compatible method that involves the simultaneous fixation of tissue and deposition of MALDI matrix and is shown to afford subcellular spatial resolution in prepared tissues. Concomitantly, the high-quality preparation and fast execution time of the presented method have the potential to improve diagnosis of malignancies involving a variety of biopsy types. EXPERIMENTAL SECTION Materials. Swiss nude male mice of 4-8 weeks of age from Charles River Laboratories (Wilmington, MA) were anesthesized and sacrificed by intraperitonal 90 mg/kg ketamine and 10 mg/ kg xylazine. All animal manipulations were carried out in the animal facility at the Brigham and Women’s Hospital in accordance with federal, local, and institutional guidelines. Tumor samples were obtained from the Brain Tumor Bank in the Department of Neurosurgery, Brigham and Women’s Hospital and analyzed under appropriate Institutional Review Board guidelines. Calibration standards and the R-cyano-4-hydroxycinnamic acid (HCCA) matrix were purchased from Bruker Daltonics (Billerica, MA). Sinapinic acid was purchased from Sigma. The pan neuronal neurofilament monoclonal antibody was from Abcam (Cambridge, MA), microtubule associated protein 2 (MAP2) polyclonal antibody from Chemicon (Billerica, MA), fluorescein (FITC)conjugated donkey anti-mouse secondary antibody was from Jackson ImmunoResearch (West Grove, PA), and Alexa 488conjugated donkey anti-rabbit secondary antibody was from Invitrogen (Carlsbad, CA). The antibody against the newly characterized protein was raised in rabbit (Agar et al., manuscript in preparation). The nucleic acid stain 4,6-diamidino-2-phenylindole (DAPI) was from Calbiochem (San Diego, CA). Chemicals (HPLC grade solvents, acids, and detergents) were from Fisher Scientific (Fair Lawn, NJ). Glass coverslips were 22 mm × 22 mm (thickness 1: 0.13-0.17 mm) from Fisher Scientific (Pittsburgh, PA). Indium tin oxide (ITO) coated coverslips with busbars were 18 mm × 18 mm with resistivity from 8 to 12 Ω (thickness 1: 0.13-0.17 mm) from SPI Supplies (West Chester, PA). ImmEdge PEN was from Vector Laboratories, Inc. (Burlingame, CA). Sample Preparation for MALDI-Mass Spectrometry Imaging. Mouse brain tissue and human tumor specimens were sectioned using a Microm HM525 cryostat from Mikron Instru(17) Lemaire, R.; Wisztorski, M.; Desmons, A.; Tabet, J. C.; Day, R.; Salzet, M.; Fournier, I. Anal. Chem. 2006, 78, 7145-7153. (18) Schwartz, S. A.; Reyzer, M. L.; Caprioli, R. M. J. Mass Spectrom. 2003, 38, 699-708.

ments Inc. (San Marcos, CA). Method optimization: In selection of a proper solvent combination for a given experiment, attention was also given to the surface tension effect on the matrix distribution. For a specimen of dimensions 16 µm thickness with a surface of 10 mm × 5 mm, the tissue volume is 0.8 µL. The ratio recommended by the Online Information Center for Immunohistochemistry IHC World (www.ihcworld.com/) being of at least 20:1, we found that using a 20 µL volume of fixative/matrix solution was optimal. We noted that the appearance of the matrix layer is a reliable indicator of its ionization/desorption efficiency potential. The most efficient matrix layers had a homogeneous and shiny appearance whereas scattered layers with a matte finish rendered lower quality spectral data. In our understanding, the ability of the matrix layer to reflect light is indicative of the matrix crystals’ homogeneity. Mass Spectrometry. A MALDI-TOF mass spectrometer Microflex interfaced with FlexImaging Software (in-house version) from Bruker Daltonics (Billerica, MA) was operated in the linear mode for the mass-to-charge ratio (m/z) greater than 3000 and in the reflectron mode for m/z less than 3000. The instrument is equipped with a standard nitrogen laser with 337 nm wavelength and a maximum repetition rate of 20 Hz. The laser spot size is estimated to be between 100 and 150 µm. Standard instrument parameters were used, although delayed extraction times were optimized to between 400 and 800 ns, depending upon the preferred mass range. To perform imaging of tissue specimens directly from coverslips in a Microflex instrument without introducing modifications of either the laser beam or molecular ions path distances, we custom-made coverslip holders from an original Bruker’s Microflex stainless steel target with a cavity of 25 mm × 25 mm and a depth of 300 µm. The holder was produced to accommodate glass coverslips of dimensions: 22 mm × 22 mm, thickness 1 (0.13-0.17 mm). The coverslips are held on the target by minimal conductive adhesive tape. Immunohistochemistry. The protocol to reduce background staining for mouse produced antibody on mouse tissue staining was all performed at room temperature, tissue sections of 9 µm thickness: blocking for 30 min with phosphate buffered saline (PBS), 2% normal donkey serum, 1% bovine serum albumin (BSA), 0.1% gelatin, 0.1% Triton X-100, 0.05% Tween-20, and 0.05% sodium azide; primary antibody for 60 min in PBS, 1% BSA, 0.01% gelatin, 0.05% sodium azide, anti-neurofilament NF 1:1000, anti-MAP2 1:1000, anti-new-protein 1:1000 (Agar et al., manuscript in preparation); rinse three times for 5 min with PBS; secondary antibody for 45 min in PBS, 2% normal donkey serum, anti-mouse FITC conjugated secondary antibody 1:200 and DAPI 1 µg/mL; wash twice with PBS for 10 min. Confocal Microscopy. Immunohistochemistry was visualized on a Leica TCS SP2 accousto optical beam splitter (AOBS) spectral confocal microscope equipped with a 405 nm UV laser. The UV laser was used to excite DAPI at 405 nm, and the 488 nm line of the argon laser was used to excite the FITC fluorophore. Samples were scanned at 0.275-0.623 µm optical slices (in the z direction). Stack images of 10-16 slices were combined using the 2D maximum projection of a series along the fixed axis option. Intensity and contrast of each acquisition channel were readjusted Analytical Chemistry, Vol. 79, No. 19, October 1, 2007

7417

for each sample using the “glow over under” option of the Leica confocal software. Electron Microscopy. Samples were fixed in a mixture of 2.5% glutaraldehyde and 2% paraformaldehyde in 0.1 M sodium cacodylate buffer (pH 7.4), washed in 0.1 M cacodylate buffer, and postfixed with a mixture of 1% osmiumtetroxide (OsO4) + 1.5% potassiumferrocyanide K4[Fe(CN)6] for 30 min (control was for 2 h), washed in water, stained in 1% aqueous uranyl acetate for 30 min (control was for 1 h), followed by dehydration in grades of alcohol (50%, 70%, 95%, 2 × 100%), and then infiltrated and embedded in TAAB Epon (Marivac Canada Inc. St. Laurent, Canada). Ultrathin sections (approximately 60-80 nm) were cut on a Reichert Ultracut-S microtome, picked up onto coppergrids, stained with 0.2% lead citrate, and examined in a “Tecnai G2 Spirit BioTWIN” transmission electron microscope. Images were taken with a 2k AMT CCD camera. RESULTS AND DISCUSSION Fixation and Matrix Deposition. The greatest obstacle in preparing tissues for MS analysis by MALDI has been the conservation of tissue integrity and protein localization upon matrix deposition. This is due to the solubilization and displacement of proteins by the commonly used 50% ACN/50% aq 0.1% TFA matrix solution (Figure 1a,b). To overcome this limitation, we have combined cold solvent tissue fixation protocols with matrix deposition. Starting from established immunohistochemistry protocols (www.ihcworld.com/), we developed a series of fixative solutions that included dissolved MALDI matrixes and used them to concomitantly fix the tissue and deposit the matrix. This resulted in a histology compatible method of tissue preparation for MSI. The perturbation of tissue morphology and protein distribution were minimal, and our optimized method even maintained some subcellular structure. Matrix crystals can be removed after imaging, providing that there is the opportunity for histology and mass spectrometry imaging of the same tissue section. This method was tested and found to be compatible with common MALDI matrixes, including HCCA, sinapinic acid, and 2,5-dihydroxybenzoic acid. Frozen tissue sections were either thaw-mounted directly onto polished or ground stainless steel targets or onto glass- and ITO-coated glass coverslips. A number of different solutions commonly used for histology fixation were found to be useable for MSF, including (a) acetone, (b) methanol, (c) 95% ethanol/5% acetic acid combination, (d) 50% methanol/ 50% acetone combination, (e) 10% glacial acetic acid/30% chloroform/60% absolute ethanol combination (Carnoy’s Fixative), and (f) 50% methanol/50% ethanol combination. Each of the aforementioned solutions was evaluated for spectral quality, reproducibility, effect on tissue morphology, and image quality. In general, solvent mixtures provided better spectra than single solvent matrix solutions. Acetone combinations, for instance, gave good spectral data but dehydrated and deformed the tissue on the micrometer scale. Washing or prefixing the tissue with ethanol was tested and found to be compatible with this MSF, although the ethanol wash resulted in a decrease in overall spectral intensity (sum of signal/noise {S/N} for all peaks) and the selective disappearance of peaks. Next, we undertook iterative optimization of the following parameters: type and ratio of solvents, fixation time and temper7418 Analytical Chemistry, Vol. 79, No. 19, October 1, 2007

Figure 1. Immunohistochemistry of mouse brain tissue to illustrate the effect of standard matrix solution application to tissue. (a,b) Cell nuclei are depicted in blue using DAPI staining, and neurofilaments are depicted in green through labeling with an FITC secondary antibody probe and visualized by confocal microscopy with a 63× oil objective; scale bars ) 24 µm. Standard fixation with paraformaldehyde is presented in part a, and a 10 min treatment with the commonly used MALDI-MS solution ACN/0.1% TFA (50:50) is presented in part b. Frozen tissue sections were of 8 µm thickness, and optical slices were ∼1 µm. (c) Confocal microscopy of matrix crystals upon mouse brain tissue sections with a 20× dry objective, scale bar ) 75 µm. The image represents a 19.3 µm optical stack of a 25 mg/mL sinapinic acid deposition by MSF. Inset displays a portion of a single optical slice (0.4 µm) to illustrate the characteristic hexagonal face of sinapinic acid crystals.

ature, drying temperature, matrix concentration, ratio of matrix solution to tissue volume, and finally tissue thickness. Preparations that did not noticeably perturb tissue morphology at the microscopic (10 µm or more) level were then tested for spectral quality. Spectral quality was judged primarily as a function of the number of detectable peaks, and then as a function of the S/N of individual peaks, such that a spectrum with 10 peaks at S/N of 10 would be considered as a higher quality than a spectrum with one peak at a S/N of 100. The following fixative solution was found to maximize both quantity and quality of analyte MS signals from tissue, while minimizing diffusion of proteins and tissue deformation, ethanol/methanol/acetonitrile/0.1% TFA (in water) in a 2:2: 1:1 ratio, containing 25-30 mg/mL of matrix. Fresh solution (less than 12 h old) and sonication19 of the matrix solutions for 30 min followed by a 2 min microcentrifugation at (15 000g) were critical for reproducible and homogeneous crystallization (Figure 1). Because of their delocalized π electron networks, matrix molecules have an inherent green fluorescence, allowing their monitoring by fluorescence confocal microscopy. Figure 1c illustrates matrix crystals upon a tissue section. To maintain a constant ratio of matrix solution to tissue volume, a surrounding physical and hydrophobic barrier approximately 3 mm away from the tissue’s (19) Lemaire, R.; Tabet, J. C.; Ducoroy, P.; Hendra, J. B.; Salzet, M.; Fournier, I. Anal. Chem. 2006, 78, 809-819.

Figure 2. Representative mass spectrum taken from a normal mouse brain tissue section of 16 µm thickness from 4000 to 20 000 m/z prepared by sinapinic acid solution fixation. The spectrum was produced from accumulation of 500 laser shots and is not smoothed or baseline corrected. The inset magnifies the region of the spectrum between 7000 and 8200 m/z to represent the data quality.

edge was delineated using an ImmEdge Pen. With a standard fixation time of 5 min, followed by a room-temperature drying time of approximately 5 min, the time required from slicing the frozen tissue to its introduction into the mass spectrometer is 10 min, producing high quality spectra, as illustrated in Figure 2. Spatial Resolution. To evaluate effects of our optimized method upon spatial resolution, we subjected treated normal mouse brain specimens to immunohistochemistry. Figure 3 compares protein distribution in tissue samples that were fixed with paraformaldehyde, acetone, or with the optimized matrix solution (ethanol/methanol/acetonitrile/0.1% TFA in water in a 2:2:1:1 ratio). Confocal microscopy of nuclei using the fluorescent stain DAPI, together with antibodies against either neurofilament (Figure 3a-d), a newly characterized soluble protein (Agar et al., manuscript in preparation) (Figure 3e-h), or the soluble protein MAP2 (Figure 3i,j) revealed intact gray matter-white matter interfaces and axon bundles and intact nuclei and neurofilaments. The sample shown in Figure 3a was fixed with paraformaldehyde as a positive control for efficient tissue fixation and presents a characteristic neuronal pattern. In comparison, samples in Figure 3b,c were, respectively, fixed with the optimized solution and acetone, both efficient solvents for MSF, and show neuronal features comparable to the paraformaldehyde control. Figure 3d displays an image of a tissue section fixed with the optimized solution in the presence of the matrix sinapinic acid. After crystallization, the matrix was removed by solubilization in methanol. The neurofilament and nuclei staining indicates that matrix crystal formation as well as crystal removal have minimal effects on tissue integrity. Samples in Figure 3e,f were both fixed with acetone and stained for nuclei and a newly characterized soluble protein. Tissue integrity is observed by a clear delineation between the two hemispheres and the upper commissure in Figure 3e and patterns of axon bundles in Figure 3f. In Figure 3g,h, the same antibody was used to illustrate preservation of a

Figure 3. Immunohistochemistry of mouse brain tissue. Images are representations of optical stacks from the whole tissue thickness, unless otherwise specified. Cell nuclei are depicted in blue using DAPI staining and different proteins are depicted in green through labeling with an FITC or Alexa 488 secondary antibody probe and visualized by confocal microscopy with a 63× oil objective; scale bars as indicated. (a-d) Comparison of the effect of different fixations on nuclei and neurofilaments integrity between standard paraformaldehyde (a), cold methanol/ethanol/acetonitrile/0.1% TFA (b), cold acetone (c), and after sinapinic acid solution fixation followed by matrix removal using methanol (d). (e-h) Distribution patterns of a newly characterized soluble protein with distinctive tissue distribution and perinuclear subcellular localization. Tissue slices were prepared using cold acetone (e, f; 1-2 µm optical slices), standard paraformaldehyde fixation (g), and our new cold methanol/ethanol/ACN/0.1% TFA solution (h). Panels i and j compare the distribution of MAP2 in mouse brain tissue fixed with paraformaldehyde (i) and with our new solution at -20 °C (j).

Analytical Chemistry, Vol. 79, No. 19, October 1, 2007

7419

Figure 4. Electron microscopy of normal mouse brain cortical tissue sections at 6800× magnification. (a, b) Control sections that have not been treated prior to standard electron microscopy preparation. Cellular ultrastructure preservation is observed. (c-f) Effect of the MSF process upon cellular ultrastructure. (c, d) Fixed with the cold ethanol/methanol/acetonitrile/0.1%TFA solution. (e, f) Fixed with cold acetone prior to standard electron microscopy preparation. The treatments appear to mostly affect cellular membranes to different degree. N indicates nuclei, M is for membrane, and NM for nuclear membrane. Scale bars ) 2 µm.

perinuclear localization by standard fixation with paraformaldehyde (Figure 3g) in comparison to fixation with our optimized solution (Figure 3h). Similarly, Figure 3i,j displays preservation of the MAP2 protein distribution upon treatment with either paraformaldehyde or our optimized solution. It was not possible to establish the MSF imposed limits of spatial resolution using confocal microscopy, although the resolution was judged to be less than 1 µm. To better estimate limitations of spatial resolution that resulted from our treatment, electron microscopy, which exhibits at least 10 nm spatial resolution, was used. After a 5 min fixation at -20 °C of 16 µm sections of normal mouse brain tissue using the presented methanol/ethanol/ acetonitrile/0.1%TFA solution (in the absence of matrix) or acetone, the tissue was chemically fixed, embedded, and further sectioned for electron microscopy. Subcellular structures were identified, including nuclei, cytoplasmic ribosomes, rough endo7420

Analytical Chemistry, Vol. 79, No. 19, October 1, 2007

plasmic reticulum, filaments, and some nuclear double membranes (Figure 4). However, some cellular membranes were partially disrupted, therefore minimizing delineation of subcellular compartments. Mitochondria, for example, were less well defined in treated samples, indicating that our method affects some subcellular structures. In regions allowing distinction, cellular membranes appeared to dilate upon solvent treatment, going from approximately 40 nm thickness to roughly 90 nm thickness. Imaging. Current tissue biopsy diagnostic procedures involve microscopic imaging to assess different tissue characteristics, such as cellular composition, morphology, proliferation, tissue vascularization, and upon availability, the distribution of specific biomarkers. Figure 5a ,b shows the most common type of histological staining, a hematoxylin and eosin (H and E) staining of respectively grade II and III meningioma samples resected from a patient with malignancy recurrence and progression after initial tumor removal. Hematoxylin stains nuclei dark blue or purple while eosin stains cytoplasm pink-orange, enabling cellular resolution of some of the aforementioned histological characteristics. Some features of grade II (Figure 5a) include prominent nucleoli, necrosis, and elevated mitotic rate (8 mitoses per 10 high power fields, a high power field being the area of the specimen visible under high magnification, 400× in this case), while the much higher mitotic rate (24-30 mitoses per 10 high power fields) observed in Figure 5b was enough to qualify as grade III. In comparison, Figure 5c,e,g shows selected molecular ions distribution over a grade II meningioma specimen, while Figure 5d,f,h displays the corresponding molecular ions over the grade III meningioma specimen from the same patient. In the MALDI-MS images of Figure 5, a given ion is represented by a single color and the color intensity represents the ion intensity at a given location (for instance, the spatial distribution of a protein of mass 15 474 ( 81 is shown in Figure 5c,d; a different ion mass is shown in Figure 5e,f). In this way, the relative molecular abundance as a function of location is illustrated. The observed spatial resolution of the MS images not only enables an assessment of tissue heterogeneity as observed from the H and E staining but also provides a direct assessment of a multitude of molecular features for comprehensive molecular diagnosis. The ability to image a sample and obtain the detailed spatial arrangement of compounds in an ordered target such as a slice of tissue with MALDI-MSI would be of enormous value in basic biological research, as well as clinical diagnosis. For example, selected ion surface maps could provide details of compound compartmentalization, site-specific metabolic processing, and selective binding domains for a very wide variety of natural and synthetic compounds. To rank a new technology in the imaging category, prepared samples must be a reasonable and ideally an exact representation of the in vivo tissue. This concept was recognized over 100 years ago and led to the development of a multitude of well-characterized histology methods for preparing tissues for microscopic study.20-22 Merging knowledge and understanding from this traditional field with MS has led us to develop the present sample preparation method to enable the routine, accurate, and comprehensive study of frozen tissue samples by MSI. Moreover, our newly developed MSF method (20) Clarke, J. L. Philos. Trans. R. Soc. London 1851, 141, 601-622. (21) Carnoy, J. B. Cellule 1887, 3, 6. (22) Blum, F. Vorlaufige Mittheilung. Z.w.M. 1893, 10, 314-315.

Figure 5. (a, b) H and E staining of a pair of progressive meningioma samples, grades II and III, respectively, representing common histology based tumor diagnosis. Light microscopy with 40× objective. The samples present heterogeneity that is observed in a higher degree of malignancy, and the arrows are showing mitotic events. Samples were fixed with paraformaldehyde, paraffin embedded, and diagnosed by a pathologist. Parts c and e and d and f display MALDI-MS molecular ion images of selected masses from a distinct pair of progressive meningioma samples, grades II and III, respectively. Scale bars ) 1 mm. The presented ions’ distributions also indicate heterogeneity, with a specific molecular signature. Specimens were prepared by sinapinic acid solution fixation (ethanol/methanol/ACN/0.1%TFA) at -20 °C for 5 min. The different colors depict selected ions of distinct m/z. Blue ) 5440 ( 29 m/z; white ) 15 474 ( 81 m/z. Two mass spectra from the grade II specimen image are shown in part g, corresponding to two distinct pixels indicated by pink and yellow boxes. The two decorated blue and white ions masses are highlighted on the spectra, demonstrating the absence of the 5440 m/z in the rich upper spectrum. The laser was rastered to every 200 µm and 100 µm on the grades II and III specimens, respectively.

provides tissue with a level of integrity that meets standard histology-pathology requirements as well as an ease of execution that makes it possible to implement in clinical settings. As the desorption and ionization processes of MALDI-MS applications involve an energy transfer from a laser source to the analyte via a crystallized compound referred to as the matrix,15 the latter must be incorporated onto the tissue sample of interest. One of the limitations preventing the broader application of MSI has been the incompatibility of MALDI sample preparations with histology requirements. This is because solvent systems and conditions routinely used to solubilize matrix compounds can solubilize biomolecules and disturb histological and cellular structures. Our new method preserves tissue structure and protein localization, and enables the preparation of images for publication in accordance with scientific guidelines for bioimaging applications.23,24 The MSF method is based upon histology protocols known to prevent proteins from moving or diffusing from their original location. It concomitantly denatures and precipitates analytes, while incorporating the matrix onto the tissue for further MSI analysis. Established histology fixation protocols exploit specific chemical action such as protein denaturation at low pH, protein precipitation induced by solvents, and limited protein diffusion at low temperatures.25 Decreasing the fixation temperature is expected to have a positive effect of minimizing diffusion during fixation, and the potentially negative effect of promoting agglomeration (easily reversible precipitation where protein denaturation does not occur). In contrast, preliminary observations indicate that fixation of smaller molecules such as peptides (