*Manuscript
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
Integrated Mass Spectrometry Imaging and Omics Workflows on the Same Tissue Section Using Grid-Aided, Parafilm-Assisted Microdissection
Jusal Quanico†, Julien Franck†, Maxence Wisztorski†, Michel Salzet† and Isabelle Fournier†*
†
Université de Lille 1, INSERM, U1192 - Laboratoire Protéomique, Réponse Inflammatoire et
Spectrométrie de Masse (PRISM), F-59000 Lille, France
Running TITLE: Integrated OMICS & MSI by grid-aided, parafilm-assisted microdissection
*Corresponding author: Prof. Isabelle Fournier, Laboratoire Protéomique, Réponse Inflammatoire et Spectrométrie de Masse (PRISM)-INSERM U1192, Bâtiment SN3, 1er étage, Université de Lille 1, Cité Scientifique, F-59655 Villeneuve d’Ascq, France, Tel : +33 (0)3 204 194, Fax : +33 (0)3 204 054, Email:
[email protected], ORCID ID: 0000-0003-1096-5044
1
ABSTRACT 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
BACKGROUNG In spite of the number of applications describing the use of MALDI MSI, one of its major drawbacks is the limited capability of identifying multi-compounds directly on the same tissue section. METHODS We demonstrate the use of grid-aided, parafilm-assisted microdissection to perform MALDI MS imaging and shotgun proteomics and metabolomics in a combined workflow and using only a single tissue section. The grid is generated by microspotting of acid dye 25 using a piezoelectric microspotter. In the gas phase, the dye is detectable as a free radical species, and its distribution can be superimposed with ion images generated by tissue components, then used as a guide to locate regions of interest and aid during manual microdissection. Subjecting the dissected pieces to the modified Folch method allows to separate the metabolic components from the proteins. The proteins can then be subjected to overnight digestion under controlled conditions to improve protein identification yields. RESULTS The proof of concept experiment on rat brain generated 162 and 140 metabolite assignments from three ROIs (cerebellum, hippocampus and midbrain/hypothalamus) in positive and negative mode, respectively, and 890, 1,303 and 1,059 unique protein accessions. CONCLUSIONS Integrated metabolite and protein overrepresentation analysis identified pathways associated with the biological functions of each ROI, most of which were not identified when looking at the protein and metabolite lists individually. GENERAL SIGNIFICANCE To the best of our knowledge, this is the first report combining both imaging and multi-omics analyses in the same workflow and on the same tissue section.
s.
Keywords: parafilm-assisted microdissection, MALDI MS imaging, metabolomics, proteomics, brain regions, multi Omics
2
INTRODUCTION 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
MALDI mass spectrometry imaging (MSI) is an MS-based technique that can be used to detect and map a wide spectrum of compounds in a single analysis [1]. MSI finds notable applications in the tracking and quantification of drugs and metabolites for pharmaceutical analysis, and in the understanding of biological processes and diseases by providing spatially-resolved molecular detail for histopathological verification [2, 3]. Coupled with statistical tools [4-6], MSI can be used to discern regions of varying phenotypes across tissue sections, and thus finds practical use for example in the identification of cancerous regions in tumor biopsies and the discrimination of different tumor stages or subtypes [7, 8]. Unlike other optical imaging methods, MSI is essentially a non-targeted technique [9], and can be used in conjunction with classical histological staining to provide comprehensive information of compounds including those that do not contain reactive functional groups such as lipids [10]. In spite of the number of applications describing the use of MSI, improvements notably in the increase of lateral resolution and detection sensitivity are ongoing. One of the major drawbacks of the technique is the limited capability of detecting and identifying compounds, notably proteins, directly on tissue. This limitation is imposed by the inherent complexity of the tissue, as the presence of abundant components in the presence of salts and biological matrix tend to induce the suppression of the signal of the less abundant ones [11]. Hence, even if identification can be derived directly from ontissue MS analysis using top-down [12] or bottom-up [13] approaches, these methods remain successful only in identifying the most abundant components. A recent systematic examination of MSI protein biomarker limits of detection addresses this problem [14]. Several approaches have been reported to circumvent this limitation. One is by correlation of MSI-generated peak lists with an MS/MS spectral database obtained from an LC MS (using a high-resolutive MS instrument) run of the extracts from a whole tissue [15] or microextracted regions [13, 16-19] . Whereas whole tissue extracts provide better compound coverage, processing of whole tissues involves extensive fractionation and purification steps and at the expense of losing spatial resolution [20]. Although microextraction, on the other hand, can provide lesser compound coverage because extraction is done on a limited region of the sample only, 3
its use can be beneficial in enabling local and in-depth examination of those specific regions by avoiding 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
protein dilution during whole tissue or organ processing [21]. In this regard, microextraction can be performed as a downstream step to further characterize ROIs defined by MSI, in addition to simply providing compound identifications for back-correlation. This approach has been demonstrated using liquid microjunction microextraction in order to compare different MSI-defined regions of interest (ROIs) in terms of their protein content [22]. Furthermore, microextractions can be performed in arraylike fashion to reconstruct compound heat maps in a similar fashion to MSI, albeit with lower spatial resolution, but with better coverage of high and low abundance proteins [23]. In line with this approach, the Parafilm-Assisted Microdissection (PAM) method was developed in order to sample specific ROIs in profiling mode [24], as well as an array of regions on a whole tissue section and generate label-freequantification-based heat maps of protein distribution based on the identification results [25]. In contrast to liquid microjunction, which utilizes a robotic system to perform microextraction in automated fashion, PAM is relatively cheap, and uses common materials found in the laboratory. Also, PAM provides better proteome coverage because proteolytic digestion can be done more efficiently and in a controlled manner (overnight in solution at 37°C, and with disulfide bond breakage and carbamidomethylation), compared to liquid microjunction and similar techniques utilising on-tissue digestion methods [26]. Equally important, it was shown in this work that MSI-generated information on compound localization (in this case, lipids) can be extended by performing spatially-directed analysis of other compound classes (proteins in this work) using direct microextraction-based sampling followed by electrospray-based MS. In spite of these advantages, it has to be noted that the developed PAM method, like the one utilizing liquid microjunction, has been applied on a consecutive tissue section, complicating comparisons and correlations because of potential artefacts introduced by the use of non-identical tissues. Furthermore, it is not often possible to obtain serial tissue sections particularly from precious tissue biopsies, hence, the importance of obtaining as much information possible from single tissue sections. Notable approaches have been reported extending information obtained from an MSI analysis by using complementary MS-based techniques as had been described by Steven and Bunch [27]. Others
4
describe interfacing with spectroscopic techniques [28]. Recently, the Bunch lab has also described the 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
use of liquid microjunction to generate heat maps of drug, lipid and protein distribution on the same tissue section in a single acquisition step by subjecting 108 points to microextraction [29], as well as the coupling of high field asymmetric waveform ion mobility spectrometry for the same purpose [30, 31]. Such approaches integrating data from analysis of multiple compound classes are invaluable in describing the processes that occur in the biological context of the tissue being analyzed. With the emergence of multiple omics strategies [32, 33], it is apparent that holistic interpretations may only be discerned upon integration of results from multiple discrete omics domains, given the complexity of biochemical pathways involving different classes of compounds [34]. The current work is reported in line with this development. Here, we aim to integrate MALDI MSI with a multiomics workflow examining the rat brain proteome and the lipidome by performing these methods on the same tissue section taking advantage of the fact that the techniques used in each step are only mildly destructive.
EXPERIMENTAL PROCEDURES Experimental Design and Statistical Rationale We first demonstrate how MS images can be acquired on different types of tissue sections mounted on parafilm-covered glass slides. For this purpose, single tissue sections were obtained from rat brain, plant root, and whole body specimens. The same rat brain section used for MS imaging was subjected to downstream lipidomic and proteomic analyses using the single-step protocol to be described in later sections. From the three different ROIs defined, PAM pieces were excised to yield protein and lipid/metabolite samples for nanoLC-MS/MS. For each PAM piece, single technical replicates were used for both lipidomic and proteomic analyses, with the lipid samples injected twice to be able to perform data acquisition in both positive and negative modes (ie., no polarity switching was employed). In the absence of additional technical replicates in this proof of concept work, qualitative differences in terms of analyte presence or absence was used to evaluate biological significance. 5
Reagents 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
HPLC grade ACN, absolute methanol (MeOH), chloroform (CHCl3) and water (H2O), and analytical reagent (AR) grade TFA were obtained from Biosolve B. V. (Dieuze, France). Acid blue 25 (45% dye content), ammonium bicarbonate (NH4HCO3), DTT, iodoacetamide (IAA), 2,5 dihydroxybenzoic acid (DHB), gelatin (type A, from porcine skin) and AR grade formic acid (FA) were obtained from Sigma-Aldrich (Saint-Quentin Fallavier, France). Carboxymethyl cellulose (CMC) was purchased from Acros Organics (Thermo Fisher Scientific, Courtabœuf, France). Sequencing grade, modified porcine trypsin was obtained from Promega (Charbonnieres, France). Sample Preparation A 12 µm longitudinal tissue section was obtained from fresh frozen rat brain using a cryostat (Leica Microsystems, Nanterre, France). This was mounted on a parafilm-covered polylysine glass slide by finger thawing. The parafilm was fixed on the slide by stretching approximately 5 cm x 10 cm parafilm and spreading it on to the slide. The parafilm was then melted by heating the slide to ensure that the parafilm does not get sucked by the vacuum system of the MALDI source. The mounted section was then dried for 15 min under vacuum, and an optical scan was taken. Roots from a live Douglas fir (Pseudotsuga menziesii) seedling were excised and rinsed five times with distilled water until the soil was completely removed. A selected specimen was embedded in 2% gelatin preheated at 40°C then solidified at 4°C. Once hardened, the specimen was stored in 20°C until use. A 10-µm longitudinal section was obtained from the specimen using the cryostat. An adult zebrafish specimen was euthanized by hypothermal shock [35] and embedded in 2% CMC. The CMC block was suspended in isopentane cooled with liquid nitrogen. A 12-µm longitudinal section was obtained using the cryostat.
Grid Printing
6
Gridlines were created on the tissue section using the CHIP 1000 Chemical Inkjet Printer 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
(Shimadzu, Japan). 10 mg/mL acid blue dye 25 was dissolved in MeOH and diluted 20x and used to mark the grid. 1000 pL of this solution was deposited on the tissue section by multipass iterative microspotting at an interval of 3 drops and waiting time of 2000 ms. Each drop contains 50 pL solution. The dwell voltage was set to 62 V and the dwell time fixed at 44.6 ms. These parameters led to 8 rounds/line deposition. 10 horizontal and 25 vertical lines were deposited 1 mm apart, leading to 1 mm2 gridlines. Matrix Deposition 15 mg/mL DHB dissolved in 70:30 MeOH/0.1% TFA solution was sprayed uniformly onto the tissue section using an improvised manual sprayer, as reported previously [25], for 25 min until matrix crystals become visible under a light microscope. High-Resolution MALDI MS Imaging Lipid images were acquired using a MALDI LTQ Orbitrap XL mass spectrometer (Thermo Fisher Scientifique, Bremen, Germany) equipped with an N2 laser and operating in positive mode and using the tissue imaging feature. The images were acquired at 60,000 mass resolution (at m/z 200) between m/z 200-1200. Four microscans were acquired per step. The automatic gain control (AGC) was turned off and the laser power was set to 70 µJ and 30 laser shots were fired per shot. The raster step size was set to 100 µm. The lipid images were processed from the MS data using ImageQuest software version 1.0.1. Lipid Extraction After MSI acquisition, the ROIs were located on the tissue section with the help of the grid coordinates. These were excised using a scalpel and immersed in 50 µL MeOH. To this solution, 25 µL CHCl3 was added and the mixture was vortexed for 1 min. 25 µL H2O was then added until an emulsion was formed, and the mixture was centrifuged for 10 min at 14,000 g. The CHCl3 layer was collected
7
and extraction was repeated using a fresh volume of CHCl3. The pooled CHCl3 layers were then dried 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
using a speedvac and stored in -20 °C until use. Protein Digestion and Extraction The aqueous layer from the previous lipid extraction steps were pooled and dried under vacuum together with the remaining parafilm pieces in the tube. The residue was then subjected to in-solution tryptic digestion as described previously [25]. Briefly, 25 µL 50 mM DTT in 50 mM NH4HCO3 was added and the solution was incubated at 55 °C for 15 min. The solution was allowed to cool before adding 25 µL 150 mM IAA in 50 mM NH4HCO3 and incubating in the dark for another 15 min. 20 µL of 20 µg/mL trypsin dissolved in 50 mM NH4HCO3 was then added and the mixture was incubated at 37 °C overnight. Formic acid was added to a total concentration of 2% (v/v) to stop the digestion, and the mixture was dried under vacuum. The precipitate was reconstituted in 20 µl 0.1 % TFA in water and subjected to desalting using C18 ziptips (Promega). The concentrated peptides were dried under vacuum and stored in -20 °C until use. Lipid Nano LC MS The extracts were reconstituted in 20 µL 50:50 CHCl3/MeOH and loaded in crimp-top vials. 5 µL was injected into a Proxeon EASY-nLC 1000 UPLC system connected to a Q-Exactive MS instrument (Thermo Fisher Scientific) and the components were eluted employing 750:250:4 MeoH/H2O/HOAc as mobile phase A and 750:250:4 ACN/MeOH/HOAc as mobile phase B. Components were separated on a 75 µm x 15 cm Acclaim PepMap RSLC column (C18, particle size 2 µm, pore size 100 Å) using an 80-min gradient (20-80% B for 20 min, 80-95% B for 20 min, 95-100% B for 5 min, 100% B for 35 min). Acquisition was done by operating in data-dependent mode. The full ms scan was set at 70,000 full width at half maximum (FWHM) resolving power (at m/z 200) in the scan range 200-1000 m/z, with an AGC target of 1e6. The top-ten method was used to select 10 precursor ions for subsequent MS/MS analysis. Separate injections were performed for the positive and negative modes of data acquisition. In positive mode, the maximum IT was set to 75 ms for the full MS scans. For the MS/MS scans, the following parameters were used: AGC target = 5e5, maximum IT = 8
60 ms, stepped normalized collision energy (NCE) = 10 V and 30 V. In negative mode, the maximum 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
IT was set to 60 ms for the full MS scans. For the MS/MS scans, the following parameters were used: AGC target = 1e6, maximum IT = 60 ms, step NCE = 10V, 20V and 35V. For both modes, the MS/MS resolution was set to 17,500 FWHM and the data-dependent settings were as follows: underfill ratio = 0.5%, Charge exclusion = unassigned and +2 or more charge states, dynamic exclusion = 8 s. Protein Ultrahigh Performance LC MS (UPLC MS) The trypsin-digested protein extracts were reconstituted with 10 µL 5% ACN/0.1% FA and injected on an EASY-nLC 1000 UPLC (Thermo Fisher Scientific) equipped with a 75 µm x 2 cm Acclaim PepMap 100 pre-column with nanoViper fittings and a 50 µm ID x 150 mm Acclaim PepMap RSLC analytical column (Thermo Fisher Scientific). The peptides were eluted using a 2-h gradient of ACN starting from 5% to 30% over 120 min at a flow rate of 300 nL/min. The Q-Exactive instrument was set to acquire top 10 MS2. The survey scans were taken at 70,000 FWHM (at m/z 400) resolving power in positive mode and using a target of 3E6 and default charge state of +2. Unassigned and +1 charge states were rejected, and dynamic exclusion was enabled for 20s. The scan range was set to 3001600 m/z. For the MS2, 1 microscan was obtained at 17,500 FWHM and isolation window of 4.0 m/z, using a scan range between 200-2000 m/z. Data Analysis Protein tandem MS/MS data were processed using MaxQuant version 1.5.3.30. Peptides were identified by using the Andromeda search engine, where target-decoy searches were performed against the Rattus norvegicus UniProt database (accessed March 4, 2014, 33,675 entries) combined with the 262 commonly detected contaminant database. The parent and fragment mass tolerances were set at 10 ppm and 20 ppm, respectively. The enzyme used was trypsin, and the maximum allowable cleavages was set to 2. Carbamidomethylation of cysteine was set as fixed modification, while oxidation of methionine and acetylation of the protein N-terminal were set as variable modifications. False discovery rates (FDR) for the peptide and protein levels were both set at 0.01. Match between runs was indicated to facilitate searches. 9
Lipid tandem MS/MS spectra were processed using Waters Progenesis QI software. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
Normalization based on Total Ion Count (TIC), alignment, and deconvolution were performed automatically with the software selecting the reference run. Peak picking was performed using default parameters and without retention time masking. Compound identifications were performed using the Progenesis Metascope algorithm (version 1.0.5673.29121) by matching spectra with the LipidMaps database (accessed March 18, 2014). The precursor and theoretical fragment ion tolerances were set at 10ppm and 20 ppm, respectively. The adducts included in the search were [M+H-H2O]+, [M+H]+, [M+Na]+ and [M+K]+ for the positive mode, and [M-H2O-H]-, [M-H]-, and M+HOAc-H]- for the negative mode. Pathway over-representation analysis using integrated data from both lipidomic and proteomic searches was performed using Integrated Molecular Pathway Level Analysis (IMPaLA) version 9 [36]. The LipidMaps identifiers were converted to the Human Metabolome Database identifiers using the Chemical Translation Service [37] available at the Fiehn Lab website (http://cts.fiehnlab.ucdavis.edu/), and these were used with the HUGO Gene Nomenclature Committee (HGNC) symbols of the proteins to perform the pathway searches projected to Homo sapiens. Q-values were obtained from Fisher’s method were set at ≤0.05.
RESULTS Development of Combined MS Imaging and OMICS Method Figure 1 shows the general scheme for the integration of the MALDI MSI workflow with shotgun MS-based metabolomics and proteomics analyses. The tissue sections were mounted on glass slides coated with parafilm prepared by melting the parafilm by heating. The images are then acquired using conventional MALDI MSI sample preparation and matrix deposition steps, and spectra are acquired using a MALDI LTQ orbitrap XL instrument. The parafilm covering of the glass slide serves as a support that allows the dissection of entire sections (as in a whole body experiment) or specific ROIs (as in localized microsampling), and at the same time does not interfere with the MALDI imaging 10
data acquisition. The dissected sections can then be subjected to stepwise lipid and protein extraction 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
using the modified Folch method, as had been demonstrated for single-sample multiomics approaches recently [32, 33, 38]. This is possible because the parafilm does not dissolve completely if immersed first in MeOH prior to the addition of CHCl3, and remains visible at the interface between the organic and aqueous layers (Figure 2). After removal of the organic layer, the remaining aqueous layer together with the parafilm piece can be dried and subjected to overnight in-solution digestion to obtain the peptides. Instances where the parafilm dissolves completely have also been observed, but the aqueous fractions nonetheless yielded peptides after overnight digestion treatment (data not shown). MALDI TOF analyses of the reconstituted organic and aqueous fractions obtained from a single PAM piece are shown in Supplementary Figure S1A-C. Full MS scan signals normally found in the region where rat brain phospholipids are detected can be observed in the organic fraction (Supplementary Figure S1A). On the other hand, signals corresponding to small proteins (Supplementary Figure S1C) and/or peptides (Supplementary Figure S1B) can be observed in the aqueous fraction relative to the control (parafilm piece not containing any tissue, spectra upside-down). The method was optimized by checking if the presence of the matrix on the tissue interferes with protein detection. As shown in Supplementary Figure S2A-B, the PAM piece from the region subjected to matrix removal yielded a lower number of protein identifications, compared to the one from the region that was not subjected to further treatment. The lipid and peptide fractions can then be subjected to shotgun LC MS for spectral acquisition, and the data processed using independent software to obtain identifications and quantitative information. These identifications can be used for high-accuracy mass correlation to obtain tentative assignments for the MS images, or in integrated pathway analyses using dedicated softwares such as IMPaLA, to generate information which can be used for further downstream analyses (such as validation of identified overexpressed candidates or functional analysis using target molecules). MS Imaging on parafilm-coated slides Whole body section sample preparation and spectral acquisition, as exemplified in this work using zebrafish sections, can be done on parafilm-covered slides as in conventional polylysine or conductive slides. It was observed that parafilm coating is advantageous especially for specimens 11
embedded in CMC, as the parafilm surface repels CMC and causes it to form droplets upon finger1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
thawing (Figure 3A). The droplets can then be wiped off from the surface so that they do not interfere in the spectra (Figure 3B). Figure 4 shows the comparison of the images obtained using the zebrafish section embedded in CMC and mounted on a glass slide alone and on another one coated with parafilm. The series of peaks corresponding to various charged states of the polymeric CMC can be easily observed on the image generated from the section mounted on the slide alone, and they interfere with the higher mass lipids observed around m/z 1100-1500. P. menziesii root tissue sections containing gelatin embedding were also mounted on parafilm-covered slides to demonstrate the applicability of this method on other types of samples. Unlike CMC which melts and gets repelled on the parafilm surface, the gelatin embedding material used for the root posed difficulties in mounting the material onto the slide. This was fixed by rapidly heating the mounted tissue so that the gelatin adheres to the parafilm surface upon melting. Using this approach, images of metabolites localized on the different regions of the root can be observed. For example, a clear distinction can be seen in the distribution of m/z 1481.300 and 1193.236 which are co-localized in the cortex, with m/z 601.491 and m/z 213.075 which are co-localized in the central cylinder (Figure 5). PAM-based Extraction of Lipid and Protein Components To perform sampling of discrete ROIs, we employed the PAM method we have previously published elsewhere [24, 25]. In contrast however to the previous method, the current approach uses the same section used for MS imaging for further downstream analysis, thus avoiding altogether the use of serial sections. To improve the localization of different ROIs and to serve as a guide for dissection, a dye grid was microspotted on the section using the CHIP 1000 instrument. Using the parameters described in the Experimental Procedures, a grid of 1 mm2 squares composed of droplets averaging 110 µm in diameter (on tissue) and 200 µm apart (from center to center) was created using acid blue 25 dye (Figure 6a). The chemical structure of the dye and its full MS spectrum generated using the MALDI LTQ Orbitrap are shown in Figure 6b and c, respectively. The observed mass of the [M]+• radical ion at m/z 394.061 is in agreement with the exact mass (m/z 394.0623). This radical ion is also observable on tissue during spectral acquisition for the MS image and can thus be used to generate the heat map 12
image of the grid (Figure 6c inset). Lipid ion images of the rat brain section subjected in this manner 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
are shown in Figure 7. After reconstruction of the lipid images, the ROIs on the rat brain section were identified and located using the grid squares. These regions, cerebellum (ROI 1), hippocampus (ROI 2) and midbrain/hypothalamus (ROI 3), were excised together with their parafilm support using a scalpel and suspended in MeOH then subjected to lipid extraction using the modified Folch method. The metabolites were separated on a C8 column and subjected to tandem MS/MS. The raw files were then interrogated in Progenesis QI by comparing the accurate precursor ion mass and MS/MS spectra with those found in the Lipid Maps curated database. Generated assignments were individually validated by counter checking the MS/MS spectra used for the assignment. Examination of the metabolites detected in positive mode yielded 162 unique metabolite identifications, 121 of which were detected in all the three ROIs (Figure 8A, Supplementary Data S1). Eight were identified only in ROI 1 and 2 each in ROI 2 and ROI 3. 14 identifications were detected in both ROI 1 and ROI 2, 7 in both ROI 2 and ROI 3, and 8 in both ROI 1 and ROI 3. In negative mode, 140 metabolites were identified, 101 of which were detected in all three ROIs (Figure 8B, Supplementary Data S1), and 7 in ROI 1 alone and 2 in ROI 2 alone. 24 were identified in both ROI 1 and ROI 2, 1 in both ROI 2 and ROI 3, and 5 in both ROI 1 and ROI 3. The proteins remaining in the aqueous phase and in the parafilm piece were subjected to overnight digestion with trypsin and the extract subjected to shotgun nanoLC-MS/MS analysis, and the protein identifications obtained after interrogation using MaxQuant. Results of the interrogation identified 890 unique protein accessions for ROI 1, 1,303 for ROI 2, and 1,059 for ROI 3 (Figure 9, Supplementary Data S2). 748 proteins were common to all three ROIs. 38 proteins were both identified in ROI 1 and ROI 2, 102 between ROI 2 and ROI 3, and 24 between ROI 1 and ROI 3. Integrated Pathway Analysis using Lipidomic and Proteomic Data To obtain the significance of the lipid and protein differential distribution across the three different ROIs, pathway over-representation analysis using the combined data was performed using the IMPaLA web tool. To be able to perform this analysis, the Lipid Maps identifiers had to be converted to the Human Metabolome database format, and the IMPaLA search projected to the human databank. It can be observed that only a small fraction of the identified metabolites could be matched with their 13
counterpart. Nonetheless, the integrated analysis yielded plenty of results when the correlation values 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
were fixed at q ≤ 0.05. The results are presented in Supplementary Data S3 and summarized in Table 1. Terms associated with neurotrophin signalling were identified, including “Neurotrophin signalling pathway” (KEGG) and “Signalling by NGF” (Reactome) both observed in ROI 1 and ROI 3, “p75 NTR receptor-mediated signalling” (Reactome) and “NGF signalling via TRKA from the plasma membrane” (Reactome) both observed exclusively in ROI3, and “TRKA receptor signalling pathway” (BioCarta) observed exclusively in ROI 2. The lipids and proteins associated with the terms are also shown in Table 1. Terms associated with G protein signalling were also found, with particular exclusivity in ROI 3 (“G protein-mediated events” (Reactome), “Inflammatory mediator regulation of TRP channels” (KEGG), “PLC beta-mediated events” (Reactome), “Diacylglycerol (DAG) and inositol trisphosphate (IP3) signalling” (Reactome), “PLC-gamma 1 signaling” (Reactome), “EGFR interacts with phospholipase C-gamma” (Reactome), “Phospholipase C-mediated cascade” (Reactome), “LPA receptor mediated events” (PID pathways) and “LPA-4 mediated signaling events” (PID pathways). Terms related to synaptic plasticity and desensitization were also noted, a portion of which were observed in both ROI 2 and ROI 3, and another portion exclusively in ROI 2. The former includes “Transmission across chemical synapses” (Reactome), “Gamma-Aminobutyric acid (GABA)ergic synapse”
(KEGG),
“Beta-agonist/beta
blocker
pathway_pharmacodynamics”
(PharmGKB),
“Neurotransmitter receptor binding & downstream transmission in the postsynaptic cell” (Reactome), “Dopaminergic synapse” (KEGG), “Trafficking of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors” (Reactome), and “Glutamate binding-activation of AMPA receptors and synaptic plasticity” (Reactome). Terms exclusively observed in ROI 2 are “Attenuation of gpcr signalling” (BioCarta), “Trafficking of GluR2-containing AMPA receptors” (Reactome), “Activation of NMDA receptor upon glutamate binding & postsynaptic events” (Reactome), “Ephrin B (EPHB)mediated forward signalling” (Reactome), and “Brain-derived neurotrophic factor (BDNF) signaling pathway” (Reactome). Terms associated with lipid biosynthesis and metabolism were also observed, although they do not associate to a specific ROI exclusively. Among these, terms “Sphingolipid de novo biosynthesis”
14
(Reactome), “Sphingolipid metabolism” (Reactome, WikiPathways and SMPD), “Metabolism of lipids 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
and lipoproteins” (Reactome), and “Inositol phosphate metabolism” (Reactome) were observed in ROI 1 and ROI 3, while “Glycosylphosphatidylinositol (GPI)-anchor biosynthesis” (EHMN) was observed exclusively in ROI 1. Whereas terms “Glycerophospholipid biosynthesis” (Reactome), “Phospholipid metabolism” (Reactome), “Glycerophospholipid metabolism” (EHMN, KEGG), “Phospholipid biosynthesis” (WikiPathways, SMPD), “Triglyceride synthesis” (WikiPathways), and “Synthesis of phosphatidylserine (PS)” (Reactome) were observed in both ROI 1 and ROI 2 only. Terms “Lipoprotein metabolism” (Reactome), “Lipid digestion, metabolism and transport” (Reactome, WikiPathway), and “Synthesis of phosphatidylcholine (PC)” were observed in all ROIs.
DISCUSSION
The PAM strategy was previously showed its efficiency to identify numerous significant number proteins (>1000) from ROIs on rat brain tissue sections as well as on prostate tumor sections. This technique is based on the manual microdissection of ROIs from the tissue according the nonsupervised segmentation experiment performed by MALDI-MSI of lipids from an adjacent section. Here, we implemented a dye-based grid to serve as a guide in co-registering the ROIs directly from the MALDI MS image to the tissue section and consequently to aid in dissecting the ROIs more properly. To achieve this, a microspotter was used to accurately deposit the dye on the tissue and create the grid. Acid blue 25 was found to be detectable in MALDI positive mode due to radical formation in the gas phase under UV irradiation. This allows the direct generation of the ion image of the radical which can then be used to create a virtual grid that can be superimposed with the ion images of the components of the tissue section. This Grid-Aided PAM (GA-PAM) strategy was used to extract and identify metabolites and proteins from the same ROIs defined by the MALDI MS image. The advantage of this approach is the possibility to perform imaging and multi-omics analysis on the same tissue section. To the best of our knowledge, this is the first report combining both imaging and multi-omics analyses in the same workflow and on the same tissue section.
15
The MALDI LTQ Orbitrap hybrid is advantageous in this workflow because of its configuration 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
[39]. The high RF voltage quadrupole assembly of the q00 in the source is separated from the sample plate using an aperture lens. As such, sample plates used in the MALDI LTQ orbitrap system do not need to be conductive. In contrast, in MALDI time-of-flight systems, conductive slides are needed because a potential difference needs to be induced in the sample plate so that the charges produced will be accelerated and enter the time-of-flight sector at a uniform injection time. This flexibility permits the use of other kinds of target surfaces not conventionally used in MALDI MS, particularly nonconductive ones such as parafilm for example. The use of nitrocellulose membranes for the imaging of imprinted avocado mesocarps on a MALDI LTQ orbitrap system has also been reported [40]. It has to be noted though that any coating of the slide should be able to withstand the vacuum pressure inside the MALDI source, as we have observed that the presence of air bubbles trapped in the parafilm coating larger than approximately 1 mm in diameter leads to the destruction of the coating and possibly loss of mounted material (Supplementary Figure S3).
The advantage of looking at pathway regulation changes at different functional levels is that the changes observed at the genetic level, for example, can be confirmed at the transcriptional or metabolic level, thus providing more evidence in support of the existence of such regulatory changes. Furthermore, the combined evidence from different functional levels could aid in identifying pathway changes that will not be observed when looking at the individual functional levels alone. From Table 1, the neurotrophic pathway involves calcium/calmodulin-dependant kinases (Beta, Delta, Gamma), the protein-tyrosine phosphatase, non-receptor type 11, 3 oncogenes (Ras-oncogene, crk sarcoma virus CT10 oncogene homolog and the v-Ha-ras harvery sarcoma viral oncogene homolog), 2 adenylate cyclase (5 and 9) and protein kinases. All these factors are linked to neurite migration via the Rho GTPases Rac1 and Cdc42. The small GTPase Ras plays an essential role in the activation of Rac1 [41]. The expression of such neurotrophic pathway in the cerebellum, hippocampus and in mid-brain is in line with previous studies [42-44]. Some differences can be observed regarding the presence or absence of certain proteins between the 3 regions but overall, no differences are registered concerning the complete neurotrophic pathway. The Protein Atlas confirms that all proteins contained in this pathway 16
are detected in the three regions, with some quantitative differences in expression across each region. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
For example, PRKCB, CRK and CAMK2B are highly expressed in the cerebellum, while CAMK2D, CAMK2G, RAP1B, ADCY5, ADCY9 and DNM2 are highly expressed in the hippocampus. For G protein signaling, a similar conclusion can be drawn due to the fact that some proteins are common to both pathways like HRAS, PRKAR2B, PRKCA, PRKCB, PRKCE, ADCY5 and ADCY9. The Synaptic plasticity & desensitization pathway is more represented in the hippocampus and mid-brain than in the cerebellum due to the fact that these two brain areas exhibit more intensive neurogenesis processes compared to cerebellum [45-47]. Nevertheless, neurogenesis also occurs in the cerebellum [48, 49]. In the cerebellum, GRM1, GRIN1, and GRIN2B are specifically detected. All 3 are related to glutamate receptor function and are implicated in cerebellum disorders [50]. Obsessive–compulsive disorder, attention deficit hyperactivity disorder, and grooming disorders are related to glutamate and its receptors [51, 52]. Glutamate-modulating agents have recently showed promise in the treatment of disorders of inhibition [50]. For signaling pathways involving oxytocin, opioids, gonadotrophin-releasing hormone (GnRH), VEGF, ghrelin, IL8, those are predominantly expressed in the hippocampus and hypothalamus and absent in the cerebellum. Oxytocin, GnRH, and ghrelin are produced by the hypothalamus and their signaling pathways are involved in the Hypothalamus-hypophysis-adrenals axis for GnRH and opioids, [53], hippocampus-hypothalamus orexin pathway for ghrelin [54] and oxytocin pathway in hypothalamus-hippocampus for learning [55]. Taken together, the results confirm that all pathways identified in the different regions are in line with the biological function and physiological processes occurring in each region.
Another advantage is that the GA-PAM offers a better depth of coverage in terms of number of metabolite and protein species detected. The lipid and protein identifications can then be used together for integrated pathway analysis. This, combined with localization information obtained from MALDI MS imaging, should provide better insight into the biological processes occuring in the ROIs examined. In the future, the GA-PAM can be applied to clinical investigations, where it should provide more information about the pathology being examined because now we can observe changes in protein and lipid expression at the same localization. 17
Acknowledgements 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
This research was supported by grants from Ministère de L’Education Nationale, de L’Enseignement Supérieur et de la Recherche (MENESR), Institut National de la Santé et de la Recherche Médicale (INSERM) and Université de Lille.
Author Contributions JQ, JF, MW and IF designed the study. JQ and JF performed all experiments. JQ, JF and MS wrote the manuscript. IF and MS provided funding for this work. All authors have read and corrected the manuscript.
Additional Information The authors declare no competing financial interests in this work. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE [56] partner repository with the dataset identifier PXD005345 (Username:
[email protected], Password: RVemJY3T).
REFERENCES [1] M. Aichler, A. Walch, MALDI Imaging mass spectrometry: current frontiers and perspectives in pathology research and practice, Lab Invest, 95 (2015) 422-431. [2] T.C. Baker, J. Han, C.H. Borchers, Recent advancements in matrix-assisted laser desorption/ionization mass spectrometry imaging, Curr Opin Biotechnol, 43 (2016) 62-69. [3] Y. Dong, B. Li, A. Aharoni, More than pictures: When MS imaging meets histology, Trends Plant Sci, 21 (2016) 686-698. [4] T. Alexandrov, MALDI imaging mass spectrometry: statistical data analysis and current computational challenges, BMC Bioinformatics, 13 Suppl 16 (2012) S11. [5] M. Galli, I. Zoppis, A. Smith, F. Magni, G. Mauri, Machine learning approaches in MALDI-MSI: clinical applications, Expert Rev Proteomics, 13 (2016) 685-696. [6] K.D. Bemis, A. Harry, L.S. Eberlin, C. Ferreira, S.M. van de Ven, P. Mallick, M. Stolowitz, O. Vitek, Cardinal: an R package for statistical analysis of mass spectrometry-based imaging experiments, Bioinformatics, 31 (2015) 2418-2420.
18
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
[7] J. Kriegsmann, M. Kriegsmann, R. Casadonte, MALDI TOF imaging mass spectrometry in clinical pathology: a valuable tool for cancer diagnostics (review), Int J Oncol, 46 (2015) 893-906. [8] M. Kriegsmann, R. Casadonte, J. Kriegsmann, H. Dienemann, P. Schirmacher, J.H. Kobarg, K. Schwamborn, A. Stenzinger, A. Warth, W. Weichert, Reliable entity subtyping in Non-small cell Lung Cancer by MALDI Imaging Mass Spectrometry on Formalin-fixed Paraffin-embedded Tissue Specimens, Mol Cell Proteomics, (2016). [9] I. Fournier, M. Wisztorski, M. Salzet, Tissue imaging using MALDI-MS: a new frontier of histopathology proteomics, Expert Rev Proteomics, 5 (2008) 413-424. [10] Y.J. Lee, D.C. Perdian, Z. Song, E.S. Yeung, B.J. Nikolau, Use of mass spectrometry for imaging metabolites in plants, Plant J, 70 (2012) 81-95. [11] L. Tomlinson, J. Fuchser, A. Futterer, M. Baumert, D.G. Hassall, A. West, P.S. Marshall, Using a single, high mass resolution mass spectrometry platform to investigate ion suppression effects observed during tissue imaging, Rapid Commun Mass Spectrom, 28 (2014) 995-1003. [12] D. Debois, N. Smargiasso, K. Demeure, D. Asakawa, T.A. Zimmerman, L. Quinton, E. De Pauw, MALDI in-source decay, from sequencing to imaging, Top Curr Chem, 331 (2013) 117-141. [13] J. Stauber, L. MacAleese, J. Franck, E. Claude, M. Snel, B.K. Kaletas, I.M. Wiel, M. Wisztorski, I. Fournier, R.M. Heeren, On-tissue protein identification and imaging by MALDI-ion mobility mass spectrometry, J Am Soc Mass Spectrom, 21 (2010) 338-347. [14] S.M. van de Ven, K.D. Bemis, K. Lau, R. Adusumilli, U. Kota, M. Stolowitz, O. Vitek, P. Mallick, S.S. Gambhir, Protein biomarkers on tissue as imaged via MALDI mass spectrometry: A systematic approach to study the limits of detection, Proteomics, 16 (2016) 1660-1669. [15] M. Lagarrigue, T. Alexandrov, G. Dieuset, A. Perrin, R. Lavigne, S. Baulac, H. Thiele, B. Martin, C. Pineau, New analysis workflow for MALDI imaging mass spectrometry: application to the discovery and identification of potential markers of childhood absence epilepsy, J Proteome Res, 11 (2012) 5453-5463. [16] Y. Schober, T. Schramm, B. Spengler, A. Rompp, Protein identification by accurate mass matrixassisted laser desorption/ionization imaging of tryptic peptides, Rapid Commun Mass Spectrom, 25 (2011) 2475-2483. [17] J.O. Gustafsson, J.S. Eddes, S. Meding, T. Koudelka, M.K. Oehler, S.R. McColl, P. Hoffmann, Internal calibrants allow high accuracy peptide matching between MALDI imaging MS and LC-MS/MS, J Proteomics, 75 (2012) 5093-5105. [18] S. Meding, K. Martin, O.J. Gustafsson, J.S. Eddes, S. Hack, M.K. Oehler, P. Hoffmann, Tryptic peptide reference data sets for MALDI imaging mass spectrometry on formalin-fixed ovarian cancer tissues, J Proteome Res, 12 (2013) 308-315. [19] L.A. McDonnell, A. Walch, M. Stoeckli, G.L. Corthals, MSiMass list: a public database of identifications for protein MALDI MS imaging, J Proteome Res, 13 (2014) 1138-1142. [20] K.L. Schey, D.M. Anderson, K.L. Rose, Spatially-directed protein identification from tissue sections by top-down LC-MS/MS with electron transfer dissociation, Anal Chem, 85 (2013) 6767-6774. [21] H. Meistermann, J.L. Norris, H.R. Aerni, D.S. Cornett, A. Friedlein, A.R. Erskine, A. Augustin, M.C. De Vera Mudry, S. Ruepp, L. Suter, H. Langen, R.M. Caprioli, A. Ducret, Biomarker discovery by imaging mass spectrometry: transthyretin is a biomarker for gentamicin-induced nephrotoxicity in rat, Mol Cell Proteomics, 5 (2006) 1876-1886. [22] J. Quanico, J. Franck, C. Dauly, K. Strupat, J. Dupuy, R. Day, M. Salzet, I. Fournier, M. Wisztorski, Development of liquid microjunction extraction strategy for improving protein identification from tissue sections, J Proteomics, 79 (2013) 200-218. [23] M. Wisztorski, A. Desmons, J. Quanico, B. Fatou, J.P. Gimeno, J. Franck, M. Salzet, I. Fournier, Spatially-resolved protein surface microsampling from tissue sections using liquid extraction surface analysis, Proteomics, 16 (2016) 1622-1632. [24] J. Quanico, J. Franck, J.P. Gimeno, R. Sabbagh, M. Salzet, R. Day, I. Fournier, Parafilm-assisted microdissection: a sampling method for mass spectrometry-based identification of differentially expressed prostate cancer protein biomarkers, Chem Commun (Camb), 51 (2015) 4564-4567. 19
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
[25] J. Franck, J. Quanico, M. Wisztorski, R. Day, M. Salzet, I. Fournier, Quantification-based mass spectrometry imaging of proteins by parafilm assisted microdissection, Anal Chem, 85 (2013) 8127-8134. [26] G.A. Harris, J.J. Nicklay, R.M. Caprioli, Localized in situ hydrogel-mediated protein digestion and extraction technique for on-tissue analysis, Anal Chem, 85 (2013) 2717-2723. [27] R.T. Steven, J. Bunch, Repeat MALDI MS imaging of a single tissue section using multiple matrices and tissue washes, Anal Bioanal Chem, 405 (2013) 4719-4728. [28] T. Bocklitz, K. Brautigam, A. Urbanek, F. Hoffmann, F. von Eggeling, G. Ernst, M. Schmitt, U. Schubert, O. Guntinas-Lichius, J. Popp, Novel workflow for combining Raman spectroscopy and MALDI-MSI for tissue based studies, Anal Bioanal Chem, 407 (2015) 7865-7873. [29] E.C. Randall, A.M. Race, H.J. Cooper, J. Bunch, MALDI Imaging of Liquid Extraction Surface Analysis Sampled Tissue, Anal Chem, 88 (2016) 8433-8440. [30] R.L. Griffiths, A.J. Creese, A.M. Race, J. Bunch, H.J. Cooper, LESA FAIMS Mass Spectrometry for the Spatial Profiling of Proteins from Tissue, Anal Chem, 88 (2016) 6758-6766. [31] J. Sarsby, R.L. Griffiths, A.M. Race, J. Bunch, E.C. Randall, A.J. Creese, H.J. Cooper, Liquid Extraction Surface Analysis Mass Spectrometry Coupled with Field Asymmetric Waveform Ion Mobility Spectrometry for Analysis of Intact Proteins from Biological Substrates, Anal Chem, 87 (2015) 6794-6800. [32] E.S. Nakayasu, C.D. Nicora, A.C. Sims, K.E. Burnum-Johnson, Y.M. Kim, J.E. Kyle, M.M. Matzke, A.K. Shukla, R.K. Chu, A.A. Schepmoes, J.M. Jacobs, R.S. Baric, B.J. Webb-Robertson, R.D. Smith, T.O. Metz, MPLEx: a Robust and Universal Protocol for Single-Sample Integrative Proteomic, Metabolomic, and Lipidomic Analyses, mSystems, 1 (2016). [33] M.A. Salem, J. Juppner, K. Bajdzienko, P. Giavalisco, Protocol: a fast, comprehensive and reproducible one-step extraction method for the rapid preparation of polar and semi-polar metabolites, lipids, proteins, starch and cell wall polymers from a single sample, Plant methods, 12 (2016) 45. [34] K. Wanichthanarak, J.F. Fahrmann, D. Grapov, Genomic, Proteomic, and Metabolomic Data Integration Strategies, Biomark Insights, 10 (2015) 1-6. [35] M. Matthews, Z.M. Varga, Anesthesia and euthanasia in zebrafish, ILAR journal, 53 (2012) 192204. [36] A. Kamburov, R. Cavill, T.M. Ebbels, R. Herwig, H.C. Keun, Integrated pathway-level analysis of transcriptomics and metabolomics data with IMPaLA, Bioinformatics, 27 (2011) 2917-2918. [37] G. Wohlgemuth, P.K. Haldiya, E. Willighagen, T. Kind, O. Fiehn, The Chemical Translation Service-a web-based tool to improve standardization of metabolomic reports, Bioinformatics, 26 (2010) 2647-2648. [38] L. Valledor, M. Escandon, M. Meijon, E. Nukarinen, M.J. Canal, W. Weckwerth, A universal protocol for the combined isolation of metabolites, DNA, long RNAs, small RNAs, and proteins from plants and microorganisms, The Plant journal : for cell and molecular biology, 79 (2014) 173180. [39] K. Strupat, V. Kovtoun, H. Bui, R. Viner, G. Stafford, S. Horning, MALDI produced ions inspected with a linear ion trap-orbitrap hybrid mass analyzer, J Am Soc Mass Spectrom, 20 (2009) 14511463. [40] P.J. Horn, C.N. James, S.K. Gidda, A. Kilaru, J.M. Dyer, R.T. Mullen, J.B. Ohlrogge, K.D. Chapman, Identification of a new class of lipid droplet-associated proteins in plants, Plant physiology, 162 (2013) 1926-1936. [41] R. Pankov, Y. Endo, S. Even-Ram, M. Araki, K. Clark, E. Cukierman, K. Matsumoto, K.M. Yamada, A Rac switch regulates random versus directionally persistent cell migration, The Journal of cell biology, 170 (2005) 793-802. [42] Z. Wei, J. Liao, F. Qi, Z. Meng, S. Pan, Evidence for the contribution of BDNF-TrkB signal strength in neurogenesis: An organotypic study, Neuroscience letters, 606 (2015) 48-52.
20
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
[43] H.J. Wang, J.P. Cao, J.K. Yu, L.C. Zhang, Z.J. Jiang, D.S. Gao, Calbindin-D28K expression induced by glial cell line-derived neurotrophic factor in substantia nigra neurons dependent on PI3K/Akt/NFkappaB signaling pathway, European journal of pharmacology, 595 (2008) 7-12. [44] L. Zhou, W. Zhou, S. Zhang, B. Liu, P. Liang, Y. Zhou, T. Zhou, K. Zhang, Y. Leng, W. Kong, BDNF signaling in the rat cerebello-vestibular pathway during vestibular compensation: BDNF signaling in vestibular compensation, The FEBS journal, 282 (2015) 3579-3591. [45] F. L'Episcopo, C. Tirolo, N. Testa, S. Caniglia, M.C. Morale, M.F. Serapide, S. Pluchino, B. Marchetti, Wnt/beta-catenin signaling is required to rescue midbrain dopaminergic progenitors and promote neurorepair in ageing mouse model of Parkinson's disease, Stem cells, 32 (2014) 21472163. [46] F. L'Episcopo, C. Tirolo, S. Caniglia, N. Testa, M.C. Morale, M.F. Serapide, S. Pluchino, B. Marchetti, Targeting Wnt signaling at the neuroimmune interface for dopaminergic neuroprotection/repair in Parkinson's disease, Journal of molecular cell biology, 6 (2014) 13-26. [47] Z. Chaker, C. George, M. Petrovska, J.B. Caron, P. Lacube, I. Caille, M. Holzenberger, Hypothalamic neurogenesis persists in the aging brain and is controlled by energy-sensing IGF-I pathway, Neurobiology of aging, 41 (2016) 64-72. [48] A. Swaminathan, M. Kumar, S. Halder Sinha, A. Schneider-Anthony, A.L. Boutillier, T.K. Kundu, Modulation of neurogenesis by targeting epigenetic enzymes using small molecules: an overview, ACS chemical neuroscience, 5 (2014) 1164-1177. [49] G. Rusanescu, J. Mao, Peripheral nerve injury induces adult brain neurogenesis and remodelling, Journal of cellular and molecular medicine, (2016). [50] M.A. Grados, E.B. Atkins, G.I. Kovacikova, E. McVicar, A selective review of glutamate pharmacological therapy in obsessive-compulsive and related disorders, Psychology research and behavior management, 8 (2015) 115-131. [51] H.B. Kim, K.M. Swanberg, H.S. Han, J.C. Kim, J.W. Kim, S. Lee, C.J. Lee, S. Maeng, T.S. Kim, J.H. Park, Prolonged stimulation with low-intensity ultrasound induces delayed increases in spontaneous hippocampal culture spiking activity, Journal of neuroscience research, (2016). [52] K.K. Ade, Y. Wan, H.C. Hamann, J.K. O'Hare, W. Guo, A. Quian, S. Kumar, S. Bhagat, R.M. Rodriguiz, W.C. Wetsel, P.J. Conn, K. Dzirasa, K.M. Huber, N. Calakos, Increased Metabotropic Glutamate Receptor 5 Signaling Underlies Obsessive-Compulsive Disorder-like Behavioral and Striatal Circuit Abnormalities in Mice, Biological psychiatry, 80 (2016) 522-533. [53] M. Salzet, D. Vieau, R. Day, Crosstalk between nervous and immune systems through the animal kingdom: focus on opioids, Trends Neurosci, 23 (2000) 550-555. [54] T.M. Hsu, J.D. Hahn, V.R. Konanur, E.E. Noble, A.N. Suarez, J. Thai, E.M. Nakamoto, S.E. Kanoski, Hippocampus ghrelin signaling mediates appetite through lateral hypothalamic orexin pathways, eLife, 4 (2015). [55] R. Feldman, M. Monakhov, M. Pratt, R.P. Ebstein, Oxytocin Pathway Genes: Evolutionary Ancient System Impacting on Human Affiliation, Sociality, and Psychopathology, Biological psychiatry, 79 (2016) 174-184. [56] J.A. Vizcaino, A. Csordas, N. Del-Toro, J.A. Dianes, J. Griss, I. Lavidas, G. Mayer, Y. Perez-Riverol, F. Reisinger, T. Ternent, Q.W. Xu, R. Wang, H. Hermjakob, 2016 update of the PRIDE database and its related tools, Nucleic acids research, (2016).
21
FIGURE LEGENDS 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
Figure 1. Integrated MSI and OMICS workflow scheme. The parafilm covering of the glass slide serves as a support that allows the dissection of entire sections or ROIs after MALDI MS image acquisition. The sections can then be subjected to downstream metabolite and protein identification by LC MS, and the data generated by both methods are used for integrated analyses. Lipid/metabolite identifications can also be correlated back to the imaging dataset to provide peak assignments. Figure 2. A) Extraction of lipids/metabolites and proteins for downstream analyses after MALDI MSI. B) The PAM piece lies on the interface of the two layers, and does not dissolve if suspended first in MeOH prior to lipid extraction. Figure 3. A) Parafilm repels the embedding material when 2% CMC-embedded tissues are mounted on parafilm-covered slides. B) Re-congealed section with the CMC droplets removed. C) Ion images of CMC-derived signals observed on zebrafish sections mounted on a polylysine glass slide (left panels) not detected on an adjacent specimen mounted on a parafilm covered slide with the CMC droplets removed (right panels). Figure 4. Pseudocolor ion images generated from MALDI MSI of whole-body zebrafish lipids mounted on parafilm-covered slide with the CMC droplets removed (A), and on a polylysine slide (B). Figure 5 : Pseudocolor ion images generated from MALDI MSI of P. menziesii root embedded in gelatin and mounted on a parafilm-covered slide, showing co-localization of m/z 247.060, 1481.300, 784.587 and 1193.236 in the root cortex, and m/z 601.491 and 213.075 in the central cylinder. Figure 6. A) 1 mm2 grid on the surface of the tissue section mounted on a parafilm-covered glass slide, created by depositing acid dye 25 using a CHIP 1000 microspotter. Insets show magnification of dye grid, and one spot with a diameter of 111 µm. Scale bars on both zooms indicate 200 µm. B) Chemical structure of acid blue 25 without the sodium adduct. C) Full MS scan showing the [M+H]+ ion of acid blue 25. Inset shows the pseudocolor ion image when m/z 394.06 is plotted.
22
Figure 7. Sample lipid ion images of the rat brain tissue section superimposed with the acid blue 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
dye signal (blue). Figure 8. Lipid and other metabolite identifications in positive (A) and negative (B) modes, for the three ROIs selected. Figure 9. Protein identifications for the three ROIs selected.
23
FIGURES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
Figure 1
Figure 2
24
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
Figure 3
25
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
Figure 4
26
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
Figure 5
27
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
Figure 6
28
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
Figure 7
Figure 8
29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
Figure 9
30
Supplementary Figures 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
Supplementary Figure S1. Full MS spectra of the dried and reconstituted organic and aqueous phases. A) Organic phase examined for phospholipid signals along the m/z 600-900 range. B) Aqueous phase examined for endogenous peptides. C) Aqueous phase examined for small intact proteins. All spectra are plotted against the control (parafilm piece not containing any tissue), shown below the axes.
31
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
Supplementary Figure S2. Comparison of number of protein identifications before and after matrix removal. A) Location of the PAM pieces in the washed (right hemisphere) and unwashed (left, with the DHB matrix visible in dark contrast) portions of the rat brain section. B) Protein identifications after subjecting the PAM pieces to extraction and nanoLC-MS/MS.
Supplementary Figure S3. Parafilm-coated slides before (left) and after (right) insertion into the MALDI source. In the slide on the left, the trapped air bubbles were removed during slide preparation.
32
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49
Table 1. Selected integrated pathway overrepresentation analysis results using combined protein and lipid/metabolite data from 3 ROIs defined following a MALDI MS imaging experiment. ROI 3
Members from Mapped Entries Proteins Lipids
ROI 1
ROI 2
(Cerebellum)
(Hippocampus)
(Midbrain/ Hypothalamus)
Neurotrophin signaling 1. Neurotrophin signaling pathway
+
-
+
2. Signaling by NGF
+
-
+
HRAS; CAMK2B; CAMK2D; CAMK2G; RAP1B; PTPN11; CRK PTPN11
3. p75 NTR receptor-mediated signaling
-
-
+
MAG
4. NGF signaling via TRKA from the plasma membrane
-
-
+
5. trka receptor signaling pathway
-
+
-
HRAS; PRKAR2B; PRKCA; PPP2CB; PRKCE; ADCY5; ADCY9; PTPN11; CRK; DNM2 HRAS; PRKCA; PRKCB
G protein signaling 1. G protein signaling pathways
-
+
+
2. Opioid signaling
+
+
+
Pathway terms
HRAS; GNG4; GNG7; PRKAR2B; PRKCA; PRKCB; PRKCE; AKAP12; PRKAR2B; ADCY5; GNA13; ADCY9; GNAL; GNA11 PPP1R1B; PPP2CB; PPP3CB; PLCB4; PRKAR2B; PRKCA; GNG4; GNG7;
Ceramide (HMDB04954); Diglyceride (HMDB07248) HMDB04954; Sphingomyelin (HMDB12095); HMDB07248 HMDB04954; HMDB12095 HMDB07248
HMDB07248
HMDB07248; 3-snPhosphatidylcholine (HMDB08354, HMDB07879, HMDB07977, HMDB08054,
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49
ADCY5; GNB4; ADCY9; GNAL
3. G protein-mediated events
-
-
+
4. Inflammatory mediator regulation of TRP channels (H. sapiens)
-
-
+
5. PLC beta-mediated events
-
-
+
6. DAG and IP3 signaling
-
-
+
7. Morphine addiction (H. sapiens)
-
+
+
8. PLC-gamma 1 signaling
-
-
+
9. EGFR interacts with phospholipase C-gamma
-
-
+
10. Phospholipase C-mediated cascade
-
-
+
PRKAR2B; PRKCA; ADCY5; ADCY9; PLCB4; GNAL PRKCA; PRKCB; PRKCE; ADCY5; CAMK2B; CAMK2D; CAMK2G; ADCY9; PLCB4; PPP1CB; PPP1CC PRKAR2B; PRKCA; ADCY9; PLCB4; ADCY5 PRKAR2B; PRKCA; ADCY9; PRKCE; ADCY5 GNG4; GNG7; SLC32A1; PRKCA; PRKCB; ADCY5; GNB4; PDE2A; ADCY9; PDE10A PRKAR2B; PRKCA; ADCY9; PRKCE; ADCY5 PRKAR2B; PRKCA; ADCY9; PRKCE; ADCY5 PRKAR2B; PRKCA; ADCY9; PRKCE; ADCY5
HMDB08649, HMDB08318, HMDB08446, HMDB08794, HMDB08610, HMDB08734, HMDB08739, HMDB08049, HMDB08123, HMDB08739, HMDB08734) HMDB07248; HMDB08054
HMDB07248
HMDB07248; HMDB08054
HMDB07248
HMDB07248
HMDB07248
HMDB07248
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49
11. Signaling pathway from G-protein families
-
+
+
12. LPA receptor mediated events
-
-
+
13. LPA-4 mediated signaling events
-
-
+
Synaptic plasticity & desensitization 1. Transmission across chemical synapses
-
+
+
2. GABAergic synapse
-
+
+
3. Beta-agonist/beta blocker pathway_pharmacodynamics
-
+
+
4. Neurotransmitter receptor binding & downstream transmission in the postsynaptic cell
-
+
+
CAMK2B; HRAS; PRKCA; PRKCB; PRKAR2B; PPP3CB HRAS; PRKCE; ADCY5; GNA13; ADCY9; CRK; GNA11 ADCY9; PRKCE; GNAL; ADCY5 SYN3; SLC6A1; GNB4; DLG1; SLC32A1; HRAS; PRKCA; GAD1; GNG7; DLG3; ADCY5; CAMK2B; CAMK2D; CAMK2G; ADCY9; CACNA2D1; GNAL; ACTN2; GNG4; PRKCB SRC; SLC6A1; GNG4; PRKCB; GNG7; SLC32A1; PRKCA; GAD1; ADCY5; GNB4; ADCY9; PLCL1 GNG4; GNG7; PRKCA; PRKCB; PRKCE; ADCY5; GNB4; ADCY9; PLCB4; SLC9A3R1; SRC HRAS; GNG4; GNG7; PRKCA; PRKCB; GRIA3; CACNG8; GRIN1; GRIN2B; GNB4; DLG1; DLG3;
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49
5. Dopaminergic synapse (H. sapiens)
-
+
+
6. Trafficking of AMPA receptors
-
+
+
7. Glutamate binding-activation of AMPA receptors & synaptic plasticity
-
+
+
8. Long-term potentiation (H. sapiens)
+
+
+
9. Attenuation of gpcr signalling
-
+
-
10. Trafficking of GluR2-containing AMPA receptors
-
+
-
11. Activation of NMDA receptor upon glutamate binding & postsynaptic events 12. EPHB-mediated forward signaling
-
+
-
-
+
-
ADCY5; CAMK2B; CAMK2D; CAMK2G; ADCY9; GNAL; ACTN2 GNG4; GNG7; GNB4; PRKCA; PRKCB; PPP2CB; ADCY5; CAMK2B; CAMK2D; CAMK2G; PLCB4; GNAL; PPP1R1B; PPP1CB; PPP1CC; GRIA3; GRIN2B; PPP3CB PRKCA; CACNG8; GRIA3; PRKCB; DLG1; CAMK2B; CAMK2D; CAMK2G PRKCB; PRKCA; DLG1; CAMK2B; CAMK2D; CAMK2G; CACNG8; GRIA3 CAMK2D; PLCB4; PPP1CB; PPP3CB; GRM1; HRAS; PRKCA; PRKCB; GRIN1; GRIN2B; CAMK2B; CAMK2G; RAP1B; PPP1CC PRKAR2B; PRKCA; PRKCB PRKCA; PRKCB; GRIA3 GRIN2B; HRAS; GRIA3; GRIN1 GRIN2B; HRAS; SRC;
HMDB07248
HMDB07248
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49
13. Brain-derived neurotrophic factor signaling pathway
-
+
-
GRIN1 HRAS; PTK2B; GRIA3; GRIN1; GRIN2B; SRC
Lipid biosynthesis & metabolism 1. Sphingolipid de novo biosynthesis
+
-
+
PPAP2B
2. Sphingolipid metabolism
+
-
+
PPAP2B; SMPD3
3. Metabolism of lipids & lipoproteins
+
-
+
ECI1; PPAP2B; SCP2; PCCB; PPP1CB; PPP1CC; P4HB; BDH1; INPP4A; CDS2; GPX1; HSD17B4; MTMR2; HSPG2; GC; ACLY; PIP4K2A; SMPD3; SACM1L; HSD17B4
4. Inositol phosphate metabolism
+
-
+
INPP5A; INPP1;
HMDB07248; HMDB04954; HMDB12095; HMDB08354, HMDB07879, HMDB07977, HMDB08054, HMDB08649, HMDB08318, HMDB08446, HMDB08794, HMDB08610, HMDB08734, HMDB08739, HMDB08049, HMDB08123 HMDB07248; HMDB04954; HMDB12095; HMDB08354, HMDB07879, HMDB07977, HMDB08054, HMDB08649, HMDB08318, HMDB08446, HMDB08794, HMDB08610, HMDB08734, HMDB08739, HMDB08049, HMDB08123 HMDB07248; L-1phosphatidylethanolamine (HMDB08999, HMDB09725, HMDB09060, HMDB08960); HMDB04954; HMDB08354, HMDB07879, HMDB07977, HMDB08054, HMDB08649, HMDB08318, HMDB08446, HMDB08794, HMDB08610, HMDB08734, HMDB08739, HMDB08049, HMDB08123; Triglyceride (HMDB05460); HMDB12095 HMDB07248
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49
ITPKA; PLCB4; INPP4A; SYNA1; NUDT3 5. Glycosylphosphatidylinositol (GPI)-anchor biosynthesis
+
-
-
6. Glycerophospholipid biosynthesis
+
+
-
CDS2
7. Phospholipid metabolism
+
+
-
CDS2; INPP4A
8. Glycerophospholipid metabolism
+
+
-
CDS2; PPAP2B
HMDB07248; HMDB08999, HMDB09725, HMDB09060, HMDB08960 HMDB07248; HMDB08999, HMDB09725, HMDB09060, HMDB08960; HMDB05460; HMDB08354, HMDB07879, HMDB07977, HMDB08054, HMDB08649, HMDB08318, HMDB08446, HMDB08794, HMDB08610, HMDB08734, HMDB08739, HMDB08049, HMDB08123 HMDB07248; HMDB08999, HMDB09725, HMDB09060, HMDB08960; HMDB05460; HMDB08354, HMDB07879, HMDB07977, HMDB08054, HMDB08649, HMDB08318, HMDB08446, HMDB08794, HMDB08610, HMDB08734, HMDB08739, HMDB08049, HMDB08123 HMDB07248; HMDB08999, HMDB09725, HMDB09060, HMDB08960; HMDB05460; HMDB08354, HMDB07879, HMDB07977, HMDB08054, HMDB08649, HMDB08318, HMDB08446, HMDB08794, HMDB08610, HMDB08734, HMDB08739, HMDB08049,
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49
9. Phospholipid biosynthesis
+
+
-
10. Triacylglyceride synthesis 11. Synthesis of PS
+ +
+ +
-
PPAP2B; GK
12. Choline metabolism in cancer
-
+
+
WASF1; HRAS; PRKCA; PRKCB; PPAP2B
13. Lipoprotein metabolism
+
+
+
P4HB; HSPG2
HMDB08123 HMDB08999, HMDB09725, HMDB09060, HMDB08960; HMDB05460; HMDB08354, HMDB07879, HMDB07977, HMDB08054, HMDB08649, HMDB08318, HMDB08446, HMDB08794, HMDB08610, HMDB08734, HMDB08739, HMDB08049, HMDB08123 HMDB07248; HMDB05460 HMDB08999, HMDB09725, HMDB09060, HMDB08960, HMDB08999, HMDB09725; HMDB08354, HMDB07879, HMDB07977, HMDB08054, HMDB08649, HMDB08318, HMDB08446, HMDB08794, HMDB08610, HMDB08734, HMDB08739, HMDB08049, HMDB08123, HMDB08318, HMDB08794 HMDB08739, HMDB08734, HMDB08649, HMDB08123, HMDB08318, HMDB08049, HMDB08794, HMDB08446, HMDB08354, HMDB07879, HMDB07977, HMDB08054; HMDB07248 HMDB07248; HMDB05460; HMDB08354, HMDB07879, HMDB07977, HMDB08054, HMDB08649, HMDB08318, HMDB08446, HMDB08794, HMDB08610, HMDB08734,
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49
14. Lipid digestion, mobilization & transport
+
+
+
15. Synthesis of PC
+
+
+
Nitric oxide signaling pathway
-
+
+
Oncostatin M signaling pathway
-
+
-
Thromboxane A2 receptor signaling
-
+
-
G alpha (z) signaling events
-
+
+
RAS signaling pathway
-
+
+
P4HB; HSPG2; PPP1CB; PPP1CC
CAMK2B; PRKAR2B; PRKCA; PRKCB; GRIN1; GRIN2B; PPP3CB HRAS; PTK2B; PRKCA; PRKCB; PRKCE; SRC PRKCA; PRKCB; PRKCE; SLC9A3R1; SRC; DNM1 PRKCA; PRKCB; GNG7; GNG4; PRKCE; ADCY5; GNB4; ADCY9 HRAS; GNG4; GNG7;
HMDB08739, HMDB08049, HMDB08123 HMDB07248; HMDB05460; HMDB08354, HMDB07879, HMDB07977, HMDB08054, HMDB08649, HMDB08318, HMDB08446, HMDB08794, HMDB08610, HMDB08734, HMDB08739, HMDB08049, HMDB08123 HMDB07248; HMDB08999, HMDB09725, HMDB09060, HMDB08960; HMDB08354, HMDB07879, HMDB07977, HMDB08054, HMDB08649, HMDB08318, HMDB08446, HMDB08794, HMDB08610, HMDB08734, HMDB08739, HMDB08049, HMDB08123
HMDB07248
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49
Gonadotrophin-releasing hormone signaling pathway
-
+
+
Oxytocin signaling pathway
-
+
+
Ca 2+ pathway
-
+
-
Calcium-dependent events
-
+
+
Leukocyte transendothelial migration (H. sapiens)
-
+
-
Transcription factor Creb & its extracellular signals
-
+
+
PRKCA; PRKCB; RASAL1; GRIN1; GRIN2B; SYNGAP1; PAK2; GNB4; RAP1B; PTPN11 PTK2B; SRC; PRKCB; HRAS; PRKCA; PRKCB; GNA11; MAP2K4; CAMK2D; CAMK2G; ADCY9; PLCB4; ADCY5; CAMK2B CACNG8; SRC; PPP3CB; HRAS; GUCY1B3; PPP1CB; PRKCA; PRKCB; CACNA2D1; ADCY5; CAMK2B; CAMK2D; CAMK2G; ADCY9; PLCB4; ACTB; PPP1CC; MYLK GNG7; PRKCA; PPP3CB; GNG4 PRKAR2B; PRKCA; ADCY9; ADCY5
EZR; PTK2B; PRKCA; PRKCB; ACTN4; CTNNA2 HRAS; PRKAR2B; PRKCA; PRKCB; PRKAR2B; CAMK2B;
HMDB07248
HMDB07248
HMDB08739, HMDB08734, HMDB08649, HMDB08123, HMDB08318, HMDB08049, HMDB08794, HMDB08446, HMDB08354, HMDB07879, HMDB07977
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49
Kit receptor
-
+
-
Ghrelin
-
+
-
IL8- & CXCR1-mediated signaling events
-
+
-
BCR signaling pathway
-
+
+
HDL-mediated lipid transport
-
+
-
Signaling events regulated by Ret tyrosine kinase
-
+
-
EGF-EGFR signaling pathway
-
+
-
CBL-mediated ligand-induced downregulation of egf receptors pathway EGF signaling pathway
-
+
-
-
+
+
VEGFA-VEGFR2 pathway
-
+
+
Signaling by VEGF
-
+
+
VEGF hypoxia & angiogenesis
-
+
-
VEGF signaling pathway
-
+
-
CAMK2D; CAMK2G HRAS; PRKCA; PRKCB; SRC; EZR PRKCA; PRKCB; SRC; PRKCE PRKCA; PRKCB; PRKCE; DNM1 HRAS; PRKCA; PRKCB; PPP3CB HMDB05460; HMDB08739, HMDB08734, HMDB08649, HMDB08123, HMDB08318, HMDB08049, HMDB08794, HMDB08446, HMDB08354, HMDB07879, HMDB07977 HRAS; PRKCA; SRC; SHANK3 HRAS; PTK2B; PRKCA; PRKCB; STMN1; ABI1; SRC; DNM1 SRC; PRKCB; PRKCA HRAS; SRC; PRKCB; PPP3CB; PRKCA HRAS; WASF1; PTK2B; PRKCA; PRKCB; ABI1; SRC HRAS; WASF1; PTK2B; PRKCA; PRKCB; ABI1; SRC HRAS; PRKCA; PRKCB; SRC HRAS; SRC; PRKCB; PPP3CB; PRKCA
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49
VEGFR2 mediated cell proliferation
-
+
-
PDGF-beta signaling pathway
-
+
+
HRAS; PRKCA; PRKCB; SRC HRAS; PRKCA; PRKCB
Supplementary Material (for online publication) Click here to download Supplementary Material (for online publication): Supplementary Data S1 LipidMetabolite Identifications.x
Compound (retention Neutral mass time_m/z) (Da) 45.01_524.4674m/z 44.88_307.1092m/z 52.99_1004.7992m/z 70.27_932.8638m/z 52.12_822.6373m/z 54.25_776.6153m/z 43.22_449.1310m/z 54.04_864.6482m/z 53.56_704.5579m/z 53.80_634.5404m/z 49.33_491.2745m/z 49.44_602.4817m/z 48.87_778.5384m/z 48.35_846.6001m/z 48.59_818.6053m/z 51.73_862.6317m/z 45.92_503.3110m/z 69.49_930.8481m/z 49.51_1011.7120m/z 50.02_367.3571m/z 42.83_697.4795m/z 60.89_668.6541m/z 21.99_637.4933m/z 60.26_898.7257m/z 59.74_856.7160m/z 36.37_215.1430m/z 81.91_760.5854m/z 67.55_636.6653m/z 61.09_676.6604m/z 60.89_843.7337m/z 60.98_856.7228m/z 58.05_457.2947m/z 40.10_591.3162m/z 58.25_450.4436n 56.14_629.4566m/z 56.36_866.6633m/z 57.40_647.6215n 38.23_417.2013m/z 58.60_540.5715m/z 68.68_928.8332m/z 72.57_792.5540m/z 51.32_748.5853m/z 48.59_718.5746m/z 51.73_720.5901m/z 51.32_438.4435n
450.4436
647.6215
438.4435
m/z
Charge
Retention time (min)
524.4674 307.1092 1004.7992 932.8638 822.6373 776.6153 449.1310 864.6482 704.5579 634.5404 491.2745 602.4817 778.5384 846.6001 818.6053 862.6317 503.3110 930.8481 1011.7120 367.3571 697.4795 668.6541 637.4933 898.7257 856.7160 215.1430 760.5854 636.6653 676.6604 843.7337 856.7228 457.2947 591.3162 451.4509 629.4566 866.6633 630.6182 417.2013 540.5715 928.8332 792.5540 748.5853 718.5746 720.5901 461.4328
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
45.0 44.9 53.0 70.3 52.1 54.3 43.2 54.0 53.6 53.8 49.3 49.4 48.9 48.4 48.6 51.7 45.9 69.5 49.5 50.0 42.8 60.9 22.0 60.3 59.7 36.4 81.9 67.6 61.1 60.9 61.0 58.1 40.1 58.2 56.1 56.4 57.4 38.2 58.6 68.7 72.6 51.3 48.6 51.7 51.3
Supplementary Material (for online publication) Click here to download Supplementary Material (for online publication): Supplementary Data S2 Protein Identifications.xlsx
LFQ intensity 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
LFQ intensity 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
LFQ intensity 4 0 0 0 0 0 20.426 20.5351 20.6572 21.0922 21.1106 21.2221 21.2871 21.3453 21.5014 21.5162 21.7143 21.7203 21.8794 21.9468 21.9607 22.0043 22.023 22.0325 22.0398 22.0645 22.1285 22.1536 22.1608 22.1781 22.2064 22.2404 22.2423 22.2706 22.345 22.3481 22.371 22.398 22.4106 22.4393 22.4947 22.533 22.5757 22.593 22.6143 22.6174 22.6835 22.7585 22.7591 22.7814
N: Peptides 2 3 7 7 2 2 2 2 2 2 3 2 3 2 3 2 2 2 3 2 4 2 14 3 2 2 3 2 2 3 2 2 2 2 3 2 2 2 3 2 2 3 2 2 2 2 4 2 2
Supplementary Material (for online publication) Click here to download Supplementary Material (for online publication): Supplementary Data S3 Significant Overrepresented Pat
Pathway Source
Number of Overlapping Genes
KEGG
5
Sphingolipid de novo biosynthesis
Reactome
1
Signal Transduction
Reactome
20
Sphingolipid metabolism
Wikipathways
0
Lipoprotein metabolism
Reactome
1
Glycerophospholipid biosynthesis
Reactome
1
Phospholipid metabolism
Reactome
2
Metabolism of lipids and lipoproteins
Reactome
11
Wikipathways
0
Reactome
1
Pathway Name
Retrograde endocannabinoid signaling Homo sapiens (human)
Lipid digestion_ mobilization_ and transport
Sphingolipid metabolism
