Human SOD1-G93A Specific Distribution Evidenced in Murine Brain

The transgenic mouse model carrying the human SOD1G93A mutant gene .... and a total of 200 laser shots were accumulated per pixel at constant laser po...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/jpr

Human SOD1-G93A Specific Distribution Evidenced in Murine Brain of a Transgenic Model for Amyotrophic Lateral Sclerosis by MALDI Imaging Mass Spectrometry Elena Acquadro,†,‡ Ilaria Caron,§ Massimo Tortarolo,§ Enrico M. Bucci,‡,∥ Caterina Bendotti,§ and Davide Corpillo*,†,‡ †

ABLE Bioscences, BioIndustry Park Silvano Fumero S.p.A., Via Ribes 5, 10010 Colleretto Giacosa, TO, Italy LIMA, BioIndustry Park Silvano Fumero S.p.A., Via Ribes 5, 10010 Colleretto Giacosa, TO, Italy § Molecular Neurobiology Laboratory, Neuroscience Department, Mario Negri Institute for Pharmacological Research, Via La Masa 19, 20156 Milan, Italy ‡

S Supporting Information *

ABSTRACT: Amyotrophic lateral sclerosis (ALS) is a progressive, fatal neurodegenerative disease caused by the degeneration of motor neurons. The transgenic mouse model carrying the human SOD1G93A mutant gene (hSOD1G93A mouse) represents one of the most reliable and widely used model of this pathology. In the present work, the innovative technique of matrix-assisted laser desorption/ionization (MALDI) imaging mass spectrometry (IMS) was applied in the study of pathological alterations at the level of small brain regions such as facial and trigeminal nuclei, which in rodents are extremely small and would be difficult to analyze with classical proteomics approaches. Comparing slices from three mice groups (transgenic hSOD1G93A, transgenic hSOD1WT, and nontransgenic, Ntg), this technique allowed us to evidence the accumulation of hSOD1G93A in the facial and trigeminal nuclei, where it generates aggregates. This phenomenon is likely to be correlated to the degeneration observed in these regions. Moreover, a statistical analysis allowed us to highlight other proteins as differentially expressed among the three mice groups analyzed. Some of them were identified by reverse-phase HPLC fractionation of extracted proteins and mass spectrometric analysis before and after trypsin digestion. In particular, the 40S ribosomal protein S19 (RPS19) was upregulated in the parenkyma and reactive glial cells in facial nuclei of hSOD1G93A mice when compared to transgenic hSOD1WT and nontransgenic ones. KEYWORDS: MALDI-TOF, imaging, amyotrophic lateral sclerosis, mouse, hSOD1G93A

■

INTRODUCTION

molecular pathways and potential therapeutic targets involved in the motor neuron degeneration of SOD1 mutant mice using proteome analysis. However, this was based mostly on the approach of analyzing tissue homogenates from the whole spinal cord or skeletal muscle.7−10 An application of the consolidated matrix-assisted laser desorption/ionization mass spectrometry (MALDI MS) technique has been developed which allows to give a biomolecular rationale to the morphological observations in a given region affected by disease in comparison to adjacent unaffected areas.11 This molecular imaging technology, called MALDI-imaging MS (MALDI-IMS), is an emerging new tool for direct in situ analysis of thin tissue sections. This technique enables, without the need for labeling and reducing the number of preparative steps, the acquisition of expression profiles with high spatial

Amyotrophic lateral sclerosis (ALS) is a progressive, fatal neurodegenerative disease caused by the degeneration of motor neurons (MNs) of the spinal cord, bulbar region, and cerebral cortex. Nowadays, no test can provide a definite diagnosis of ALS, and no cure has yet been found.1 Mutations in the gene that produces the Cu/Zn superoxide dismutase 1 (SOD1) enzyme are associated with approximately 20% cases of familial ALS (fALS), accounting for 5−10% of all cases.2 The transgenic mouse model carrying the human SOD1G93A mutant gene (hSOD1G93A mouse) represents the most reliable and widely used model of the fALS,3−6 well characterized for the spinal and bulbar MNs degeneration. This model shows a similarity with ALS patients in respect to neuropathological features such as motor neuron loss, protein inclusions, and reactive gliosis in damaged regions.6 However, also in the presence of this certain genetic trigger, the cause of the selective motor neuron loss is not known. Much effort has been put, so far, to identify © 2014 American Chemical Society

Received: August 31, 2012 Published: February 28, 2014 1800

dx.doi.org/10.1021/pr400942n | J. Proteome Res. 2014, 13, 1800−1809

Journal of Proteome Research

Article

as controls, all 24 weeks old. The mice were sacrified by decapitation or intracardiac perfusion under deep anesthesia, and the brain tissues were removed and immediately frozen in isopentane at −40 °C for 3 min and stored at −80 °C until use. All the procedures involving animals and their care were carried out in conformity with the institutional guidelines of ‘‘Mario Negri’’ institute (IRFMN) that are in compliance with national (D.L. No. 116, G.U. Suppl. 40, Feb. 18, 1992, Circular No. 8, G.U., July 14, 1994) and international laws and policies (EEC Council Directive 86/609, OJ L 358, 1, Dec. 12, 1987; NIH Guide for the Care and use of Laboratory Animals, U.S. National Research Council, 1996). All the experiments and the protocol proposed were examined and approved first by Institutional Ethical Committee and then sent to the Italian Ministry of Health for authorization. The mice were bred and maintained in a SPF environment. Animals with substantial motor impairment had food on the cage bottom and water bottles with long drinking spouts. The survival time was defined as the time when the animals were unable to right themselves within 10 s after being placed on either side. At this point, the animals were deeply anesthetized with Equithesin and then sacrificed by decapitation before proceeding to the dissection of tissues for biochemical analyses.

resolution while preserving topographical information about molecular distributions and localization.12,13 Spatial relationships of molecules within a specimen are preserved because intact tissue is directly analyzed without homogenization. In this way, molecules can be interrogated in their native environments, providing new insights into the biological processes involved.14 Therefore, MALDI-IMS has several advantages in comparison to other traditional techniques: high sensitivity, high specificity, no needs for labeling, and the simultaneous detection of many analytes at a time, without prior knowledge of their identities and without the need for target-specific reagents such as antibodies. It has been shown to be amenable for the analysis of proteins, peptides (both endogenous and enzymatically produced), lipids, and small molecules (such as drugs and endogenous metabolites).15−19 The emergence of MALDI-IMS has provided an enormous impulse to the field of functional imaging applied to the study of the central nervous system (CNS). Direct tissue analysis by MALDI-IMS is an important technology for assessing the localization of molecular species and for revealing the underlying molecular signatures indicative of disease. Over the past decade, MALDI-IMS has been increasingly employed, particularly in the field of neuroscience, to assay molecular profiles directly from tissue slices. MALDI-IMS has been used to compare molecular patterns in prelesion and postlesion areas in different animal models of progressive Parkinson’s disease20 as well as to evaluate amyloid β peptide distribution in brain sections of Alzheimer’s disease mouse model.21 To the best of our knowledge, the only specific application of this technique in the study of ALS is a study concerning the analysis of post-mortem human spinal cord samples.22 On the basis of these considerations, we decided to apply MALDI-IMS in the study of the most widely used model of familial ALS, the hSOD1G93A transgenic mice, at different stages of the disease. In particular, we examined the brainstem regions such as the facial and trigeminal motor nuclei that are affected in these mice during the disease progression23 but are so small that would be difficult to analyze with classical proteomics approaches. In the present study, MALDI-IMS was used to analyze the protein profiling in brainstem sections from three mice groupstransgenic hSOD1G93A, transgenic hSOD1WT (carrying the human SOD1 wild-type gene) and nontransgenicin order to possibly identify changes in proteins abundance and/or distribution relevant to the pathological alterations of these regions.

■

Tissue Section Preparation

Experiments were performed on 10−12 μm sections of fresh frozen mouse brain at the level of brainstem and cerebellum. Tissues were cut in a cryostat (Leica Microsystems) and thawmounted onto an indium−tin oxide (ITO)-coated glass slides (Bruker Daltonics) for MALDI-IMS. Two additional serial sections were mounted on histological glasses for Hematoxylin/Eosin or immunohistochemical staining. Immediately after sectioning, the ITO-coated slices were washed and fixed with 70% and 95% (v/v) ethanol for 1 min each and desiccated in vacuum for at least 1 h before matrix deposition. The MALDI matrix solution (20 mg/mL sinapinic acid in 50% (v/v) acetonitrile and 0.2% (v/v) trifluoroacetic acid) was spray-coated using a pneumatic thin-layer chromatography (TLC) sprayer (Sigma Aldrich) with a constant (0.5 bar) nitrogen flow. In order to test reproducibility of matrix coating, after around 20 spray cycles, the homogeneity of crystallized matrix layer was verified by an optic microscope. MALDI-IMS Analysis and Image Processing

The MALDI-IMS analyses were performed on an Ultraflex II time-of-flight mass spectrometer (Bruker Daltonics) with a SmartBeam laser operating at 200 Hz in positive linear mode using FlexControl 3.0 and FlexImaging 2.1 software packages (Bruker Daltonics). Ions were detected in the 2000−25 000 m/ z mass range with a sampling rate of 0.1 GS/s. The lateral resolution for MALDI-IMS was set to 80 or 90 μm, and a total of 200 laser shots were accumulated per pixel at constant laser power. A standard protein mixture (Bruker Daltonics) was employed for spectra calibration, which was done externally on the same target before each measurement. Two-dimensional ion intensity maps and average spectra were created by FlexImaging 2.1 software (Bruker Daltonics).

MATERIALS AND METHODS

Animals

Transgenic mice expressing mutant human SOD1 (hSOD1G93A, originally obtained from Jackson Laboratories, Bar Harbor, U.S.A.), crossbred with a C57BL/6 mice strain,3 were chosen as the ALS model. For the experiments, three female mice at the age of 10 weeks (presymptomatic), three at 16 weeks (first detectable neuromuscular deficit) and three at 24 weeks (end stage, with impaired paw grip strength and rotarod test performance) were used. The symptoms of neuromuscular deficit (reduced extension reflex of hind limbs when raised by the tail, reduced grip strength, impairment on rotarod) and body weight were assessed once a week. Three nontransgenic littermates (Ntg) and three transgenic mice overexpressing the human SOD1WT (hSOD1WT) were used

Statistical Analyses

On the basis of the anatomy and histological staining, regions of interest (ROIs) were defined in the tissue section using the FlexImaging 2.1 software (Bruker Daltonics), in order to select spectra associated to the nuclei to be analyzed. ROIs with a comparable size were created, therefore containing the same 1801

dx.doi.org/10.1021/pr400942n | J. Proteome Res. 2014, 13, 1800−1809

Journal of Proteome Research

Article

number of spectra (around 40). In particular, for the facial nuclei and the cerebellum three classes of spectra were created: 24 week old hSOD1G93A, hSOD1WT, and Ntg. Statistical analyses were carried out using the ClinProTools 2.2 software (Bruker Daltonics) and, during the preprocessing steps, the extracted mass spectra were internally recalibrated on common “background” peaks (this process is also known as spectral alignment) with a maximum peak shift of 0.2%, normalized on their total ion count in the observation mass range and corrected by a baseline subtraction, using a convex hull algorithm (baseline flatness 0.80), with the exclusion of null or not recalibrated spectra. An average spectrum created for each class was used for peak picking and to define integration ranges to be used to obtain the area of each peak (m/z species) in each single spectrum. Peak areas used for calculation and statistical analysis were identified by pairwise comparison using the zero level integration type and a nonparamentric Wilcoxon/ Kruskal−Wallis test was performed, with a significance cutoff point of p-value less than 0.01. A receiver operator characteristic (ROC) analysis was performed to highlight peaks discriminating the compared classes (criterion: area under curve (AUC) > 0.80).

only allowed modification was methionine oxidation. Acceptance identification criteria were the following: at least three unique peptides and a probability-based Mowse score higher than 47 (p < 0.05). Western Blotting

Total proteins were extracted from 10 frozen sections (12 μm thickness, adjacent to those analyzed by MALDI-IMS) in 100 μL of lysis buffer (1% w/v sodium dodecyl sulfate (SDS), 120 mM NaCl, 50 mM Tris-HCl pH 6.8) supplemented with protease inhibitors. After homogenization, samples were kept on ice for 15−20 min and, lastly, centrifuged at 16 000g for 10 min at 4 °C. The supernatant was recovered and stored at −20 °C until analysis. Total proteins were quantified by the bicinchoninic acid method (Sigma). Fifty micrograms of total proteins were separated on a 10% (w/v) SDS-PAGE gel and transferred by the semidry transfer method to nitrocellulose membrane (GE Healtchare). The membrane was blocked with 5% (w/v) nonfat milk in Tris-buffered saline (TBS) containing 0.1% (v/v) Tween 20 (TBST) for 1 h at 42 °C and then incubated overnight at 4 °C with rabbit polyclonal antibody to SOD1 (sc.11407, Santa Cruz Biotechnology, recognizing both human and murine SOD1) diluted 1:200 in TBST containing 1% (w/ v) nonfat milk. Blot was then washed four times in TBST, incubated 2 h at room temperature with a goat anti rabbit immunoglogulin G conjugated with horseradish peroxidase HRP (sc-2004, Santa Cruz) diluted 1:10 000 in TBST, 1% (w/ v) nonfat milk, washed two times in TBST, one time in TBS, and lastly identified on the X-ray Film by development of HRPmediated reaction using enhanced chemiluminescence (ECL) Western Blotting substrate kit (ECL Pierce, ThermoFisher).

Protein Extraction and Identification

In order to identify peaks of interest, proteins from 10 sections, contiguous to those analyzed by MALDI-IMS, were extracted with 1 M HCl by sonication in ice and overnight incubation at 4 °C. After centrifugation at 16 000g for 30 min at 4 °C, the supernatant (200 μL) was loaded onto a 4.6 × 250 mm C18reverse-phase (RP) HPLC Atlantis column (Waters) and fractionated by a 30 min linear gradient elution from 25 to 50% (v/v) acetonitrile (containing 0.1% (v/v) trifluoroacetic acid) at a flow rate of 1 mL/min by a 1525 HPLC system (Waters). Twenty fractions (1.5 mL each) were collected, covering the entire gradient period. Each fraction was evaporated and resuspended with 50 μL of 50% (v/v) acetonitrile, and then a small aliquot was analyzed by MALDI-TOF mass spectrometry (Bruker Daltonics) in order to detect protein peaks content. Subsequently, samples were again evaporated and digested (0.02 mg/mL sequence grade-modified trypsin, Promega, in 15 μL of 25 mM ammonium bicarbonate) overnight at 37 °C. One microliter of each digested sample was loaded onto a ZORBAX 300 SB C18 RP column (Agilent) and eluted with a 30 min linear gradient of acetonitrile from 5 to 70% (v/v, containing 0.1% (v/v) formic acid) at a flow rate of 0.3 μL/min by a HP1100 nanoHPLC system coupled to a XCT-Plus nanosprayion trap mass spectrometer (Agilent). MS parameters were the following: scan range m/z = 100−2200, scan speed = 8100 m/z s−1, dry gas flow = 5 L/min, dry temperature = 300 °C, capillary = 1.8 kV, skimmer = 40 V, ion charge control target = 125 000, and maximum accumulation time = 300 ms. Positively charged peptides ions were automatically isolated and fragmented; in particular, the four most intense ions of each scan were isolated and then fragmented by a voltage ramping from 30 to 200% of set fragmentation amplitude (1.30 V). Spectra were then deconvoluted by the DataAnalysis 3.4 software (Bruker Daltonics, Bremen, Germany). Protein identification was performed by MASCOT MS/MS Ion earch engine (www.matrixscience.com) using NCBInr database 20140122 filtered for the taxonomy “mus musculus” or “human”. Mass tolerance for the monoisotopic peak masses was set to 1.2 Da (parent ion) or 0.6 Da (fragments), whereas the maximum number of missed cleavages was set to two. The

Immunohistochemistry for SOD1

Immunostaining of frozen slices was done by the avidin−biotin peroxidase complex method using a Vectastain Elite ABC kit (Vector Laboratories). Briefly, after postfixation with 4% (w/v) paraformaldehyde in phosphate-buffered saline (PBS), tissue sections were incubated in 0.5% (v/v) hydrogen peroxide to block endogenous peroxide activity for 15 min and then in normal goat serum to block nonspecific binding, for 45 min at room temperature. Sections were then incubated with rabbit polyclonal antibody to SOD1 (sc.11407, Santa Cruz Biotechnology) diluted 1:200 in PBS containing 1% (w/v) bovine serum albumin and 0.1% (v/v) Triton X-100 for 1 h at room temperature. Tissue sections were then treated with biotinylated goat antirabbit immunoglobulin G (Vector Laboratories) diluted 1:250 in PBS containing 0.1% (v/v) Tween 20 (PBST) for about 1 h at room temperature, followed by treatment with ABC complex diluted 1:100 in PBST and stained with 3,3′diaminobenzidine (DAB) substrate kit (Vector Laboratories) according to the supplier’s protocol. Immunohistochemistry for 40S Ribosomal Protein S19 (RPS19)

Because the antibody to RPS19 did not work properly in frozen postfixed sections, two transgenic hSOD1G93A, two transgenic hSOD1WT, and two Ntg mice were anesthetized with Equithesin and transcardially perfused with 50 mL of PBS, followed by 50 mL of 4% (v/v) paraformaldehyde solution in PBS. Brains were rapidly removed, postfixed overnight in fixative, transferred to 30% (w/v) sucrose solution in PBS until they sank, frozen in N-pentane at −45 °C, and conserved at −80 °C. 1802

dx.doi.org/10.1021/pr400942n | J. Proteome Res. 2014, 13, 1800−1809

Journal of Proteome Research

Article

Tissues sections (20 μm) were cut on a cryostat at −20 °C and processed for immunostaining with RPS19 antibody (Proteintech Europe) using the avidin−biotin peroxidase complex method (Vectastain Elite ABC kit, Vector Laboratories). Briefly, floating coronal sections containing the facial nuclei were incubated in 1% (v/v) hydrogen peroxide to block endogenous peroxide activity for 10 min and then with 10% (v/ v) normal goat serum and 0.1% (v/v) Triton X-100 in PBS to block nonspecific binding, for 1 h at room temperature. Sections were then incubated overnight at room temperature with RPS19 antibody (Proteintech Europe) diluted 1:100 in PBS containing 1% (v/v) normal goat serum and 0.1% (v/v) Triton X-100. Tissue sections were then treated with biotinylated goat antirabbit immunoglobulin G (Vector Laboratories) diluted 1:200 in PBS containing 1% (v/v) normal goat serum for 1 h at room temperature, followed by treatment with ABC complex diluted 1:100 in PBS and stained with DAB substrate kit (Vector Laboratories) according to the supplier’s protocol.

■

RESULTS In this study, MALDI imaging MS experiments were performed to compare protein expression profile and distribution in brain sections of three mice groups (transgenic hSOD1G93A, transgenic hSOD1WT, and nontransgenic), the final goal being to differentiate the effects due to the presence of the transgenic protein from those due to the presence of the mutation. For hSOD1G93A, three ages were chosen: 10 weeks (presymptomatic), 16 weeks (first detectable neuromuscular deficit), and 24 weeks (end stage). Coronal sections of the regions comprising the facial nuclei (points −5.88 to −6.36 from bregma) and the motor trigeminal nuclei (points −4.96 to −5.34 from bregma)24 were selected for analysis, as it has been reported that these nuclei undergo neurodegeneration during ALS progression.23,25−27 First of all, SOD1 expression in 24 week old transgenic mice was verified by Western blot analysis of proteins extracted from the brainstem region: a high and comparable expression of human SOD1 in both hSOD1G93A and hSOD1WT mice was found, if compared to the smaller amount of endogenous murine SOD1 detected in Ntg mouse (see Supporting Information). Then, entire sections of the region comprising the facial nuclei and those including the trigeminal nuclei were acquired and analyzed by MALDI-IMS. The first result of this analysis performed on 24 week old mice was the appearance, in both transgenic mice groups, of a peak (indicated by the arrows in Figure 1) which was not present in Ntg mice. Its average m/z value (calculated considering both groups) was estimated to be 15 885. Although in hSOD1WT mice this peak is uniformly distributed in the sections examined, in hSOD1G93A mice it appears strongly and specifically concentrated in the facial and in the trigeminal nuclei (Figure 1, right, and Figure 2). The identity of this peak as SOD1 was confirmed through a complex procedure that combined LC intact protein fractionation followed by MALDI-TOF MS analysis and LC-MS/MS after tryptic digestion (as reported in more detail hereafter). Its accumulation in the described nuclei was further investigated by immunohistochemistry (Figure 3, and comparison between MALDI-IMS and immunohistochemistry images for hSOD1G93A is shown in Supporting Information), which evidenced the presence of SOD1 aggregates in hSOD1G93A mice, together with a high extent of vacuolization and tissue

Figure 1. MALDI-IMS average spectra of coronal brain sections in the region comprising facial nuclei from 24 week old hSOD1G93A (top), hSOD1WT (middle), and Ntg (bottom) mice. Left: entire section (delimited in red in the Hematoxylin/Eosin-stained section); right: only facial nuclei (delimited in cyano in the Hematoxylin/Eosinstained section). The same behavior was observed in trigeminal nuclei (data not shown). Arrows indicate the peak at 15 885 m/z.

degeneration. A similar phenomenon was already reported in the spinal cord.28,29 With the aim to evaluate when this accumulation begins, the same analyses were performed on sections derived from hSOD1G93A mice at different stages of the disease. In this way, it was evidenced both by MALDI-IMS investigation (Figure 4) and by immunohistochemistry (Figure 5) that SOD1 accumulation in facial and trigeminal nuclei is not detectable in the presymptomatic phase (10 week old), but it appears when the initial neuromuscular deficit becomes detectable (16 week old) and remarkably increases at the end stage (24 week old). Afterward, in order to understand if this phenomenon is strictly dependent on the presence of the mutation, brain sections of hSOD1WT mice that began to show some neuromuscular deficit (37 weeks) were analyzed, but in this case, no accumulation was found (data not shown). To better characterize proteome differences among the three 24 week old mice groups, a statistical analysis on the spectra obtained by MALDI-IMS experiments was performed by ClinProTool software, in order to find other proteins than SOD1 resulting as differentially expressed among these groups. This analysis was focused on the facial nuclei region, which showed more SOD1 accumulation in respect to the trigeminal 1803

dx.doi.org/10.1021/pr400942n | J. Proteome Res. 2014, 13, 1800−1809

Journal of Proteome Research

Article

Figure 2. Peak at 15 885 m/z distribution by MALDI-IMS in coronal brain sections from 24 week old hSOD1G93A (panel A), hSOD1WT (panel B), and Ntg mice (panel C) in the facial (bregma −6.00, left) and trigeminal (bregma −5.02, right) nuclei region. The corresponding map representations (from Paxinos and Franklin,24 with permission) are shown on the top, where facial nuclei are highlighted in red and trigeminal nuclei in green. Bar indicates 2 mm.

Figure 3. SOD1 distribution by immunohistochemistry in coronal brain sections from 24 week old hSOD1G93A (panels A and B), hSOD1WT (panels C and D), and Ntg mice (panels E and F) in the facial (left) and trigeminal (right) nuclei region. Bar indicates 60 μm. The chosen antibody recognized both human and murine SOD1.

nuclei region (see Figure 2), using the cerebellum cortex present in the same sections as control area, as no evidence of degeneration in this region has been reported so far. ROIs of facial nuclei and cerebellum cortex were carefully defined in every MALDI-IMS experiment, and calculated peak areas were then submitted to statistical analysis, as described in Methods. All three comparisons were considered: hSOD1G93A versus hSOD1WT, hSOD1G93A versus Ntg, and hSOD1WT versus Ntg. Moreover, a ROC analysis was performed, enabling us to highlight peaks discriminating the compared classes (criterion: AUC > 0.80). Results are summarized in Table 1. In particular, behavior of the peak at 16 079 m/z was found to correlate with that of SOD1, as they were markedly upregulated in facial nuclei of hSOD1G93A mice. In order to confirm the identity of the peak at 15 885 m/z and try to identify the other peaks present in the average spectra (and in particular those evidenced by the statistical analysis), a protocol involving protein extraction, C18-RP HPLC fractionation, MALDI-TOF mass spectrometric analysis, trypsin digestion, nanoLC-nanospray-ion trap mass spectrometry analysis, and database search was optimized. Fractions deriving from six independent extractions were analyzed, and hundreds of proteins were identified. However, a rigorous selection was performed: first of all, the identified protein should have a theoretical mass corresponding to one of the peaks found in the MALDI-TOF spectrum of the corresponding HPLC fraction. Moreover, this peak should be detectable in MALDI-IMS experiments. Lastly, the same identification should occur in the same fraction of at least three different extractions. This protocol allowed us to successfully identify 16 peaks (Table 2 and Supporting Information), highlightening

some common proteins previously identified by similar approaches (e.g., histones, ubiquitin, and thymosin).20,30,31 In particular, the peak at 15 885 m/z was confirmed to be SOD1, whereas the peak at 16 079 m/z was identified as 40S ribosomal protein S19-RPS19 (see Supporting Information). The latter result was verified by immunohistochemistry (Figure 6), which showed a more diffused RPS19 staining in the facial nucleus of hSOD1G93A mice (compared to hSOD1WT and Ntg). In addition, small cells, probably reactive glial cells, appeared intensely immunostained in this region. The sum of these contributions could have compensated for the loss of signal due to the motor neuron death at the end stage of the pathology. The other proteins evidenced by the statistical analysis were identified as acyl-CoA binding protein (9936 m/z, found overexpressed in the hSOD1G93A versus hSOD1WT facial nuclei comparison) and ubiquitin (8565 m/z, found downregulated in the hSOD1G93A versus Ntg facial nuclei comparison); however, the peak at 4810 m/z (found downregulated in both the hSOD1G93A versus Ntg and the hSOD1WT versus Ntg cerebellum comparisons) was not identified. These results still remain to be validated.

■

DISCUSSION The combination of MALDI-IMS with the analysis of brain slices of animal models is getting increasing attention in the field of neurodegenerative diseases, as the identification of alterations in protein or pharmaceutical tissue distribution should provide relevant information on the specific area of the 1804

dx.doi.org/10.1021/pr400942n | J. Proteome Res. 2014, 13, 1800−1809

Journal of Proteome Research

Article

Figure 4. Peak at 15 885 m/z distribution by MALDI-IMS in coronal brain sections from 10 (panel A), 16 (panel B), and 24 week old (panel C) hSOD1G93A mice in the facial (bregma −6.00, left) and trigeminal (bregma −5.02, right) nuclei region. The corresponding map representations (from Paxinos and Franklin,24 with permission) are shown on the top, where facial nuclei are highlighted in red and trigeminal nuclei in green. Bar indicates 2 mm.

Figure 5. SOD1 distribution by immunohistochemistry in coronal brain sections from 10 (panels A and B), 16 (panels C and D), and 24 week old hSOD1G93A mice (panels E and F) in the facial (left) and trigeminal (right) nuclei region. Bar indicates 50 μm. The chosen antibody recognized both human and murine SOD1.

CNS giving further insights about new therapeutic targets for human pathologies.32 To date, the extreme malleability of proteomics has allowed the unraveling of primary events in the pathogenesis of ALS,10,33 to investigate interactions of specific target such as TAR-DNA binding protein 43 (TDP-43)34 as well as to identify specific protein modifications as molecular targets of oxidative stress.35 In the present study, results from MALDI-IMS and immunohistochemistry analyses evidenced that transgenic SOD1, even if expressed at similar levels in hSOD1WT and hSOD1G93A mouse brain, has a totally different localization in these two models. Although in the former this protein is uniformly distributed along the analyzed sections, in the latter it accumulates especially in the facial and trigeminal nuclei, where it tends to form aggregates; this pattern perfectly correlates with disease progression. These results were confirmed by a statistical analysis. In particular, in the facial nuclei, the SOD1 peak contributed to discriminate hSOD1G93A from both hSOD1WT and Ntg mice, but it did not contribute in discriminating hSOD1WT from Ntg mice. These results confirm that in these nuclei SOD1 accumulation occurs only in hSOD1G93A mice and that hSOD1WT and Ntg mice have no relevant difference in SOD1 content. On the contrary, in the cerebellum, the SOD1 peak contributed to the discrimination of both hSOD1G93A and hSOD1WT mice from Ntg, whereas it did not contribute to discriminating the two transgenic mice, confirming that in this region there is no difference in SOD1 distribution and the only detectable difference is the overexpression of transgenic SOD1

Table 1. Statistical Comparisona hSOD1G93A vs hSOD1WT region

m/z

AUC

av G93A

av WT

p-value

facial nuclei

15 885 16 079 9936

0.97 0.92 0.81

700 92 100

81 30 55