MALDI MSI of Lipids in Experimental Model of ... - ACS Publications

an experimental rat model of moderate controlled cortical impact (CCI) while considering ..... same location on the right parietal cortex using an imp...
0 downloads 0 Views 1MB Size
Download PDF
Subscriber access provided by Nottingham Trent University

Article

MALDI MSI of Lipids in Experimental Model of Traumatic Brain Injury Detects Acylcarnitines as Injury Related Markers Khalil Mallah, Jusal Quanico, Antonella Raffo-Romero, Tristan Cardon, Soulaimane Aboulouard, David Devos, Firas Kobeissy, Kazem Zibara, Michel Salzet, and Isabelle Fournier Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.9b02633 • Publication Date (Web): 14 Aug 2019 Downloaded from pubs.acs.org on August 15, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 9 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

Analytical Chemistry

MALDI MSI of Lipids in Experimental Model of Traumatic Brain Injury Detects Acylcarnitines as Injury Related Markers Khalil Mallah1,2, Jusal Quanico1, Antonella Raffo-Romero1, Tristan Cardon1, Soulaimane Aboulouard1, David Devos3, Firas Kobeissy4, Kazem Zibara2, Michel Salzet1*, Isabelle Fournier1* 1Université de Lille, INSERM, U1192 - Laboratoire Protéomique, Réponse Inflammatoire et Spectrométrie de Masse (PRISM),

F-59000 Lille, France 2ER045, PRASE, Laboratory of Stem Cells, Department of Biology, Faculty of Sciences-I, Lebanese University, Beirut, Lebanon 3Department of Neurology, Expert center for Parkinson’s disease, Department of Pharmacology, University of Lille, CHU LILLE, INSERM UMR_S 1171, LICEND, France 4Department of Biochemistry and Molecular Genetics, Faculty of Medicine, American University of Beirut, Beirut, Lebanon ABSTRACT: Identifying new lipid markers linked to traumatic brain injury (TBI) is of major importance in characterizing their central role in the regeneration process and inflammatory response in such an injury model. In the present study, an advanced lipidomics analysis using high spectral resolution MALDI Mass Spectrometry Imaging was performed on different brain regions in an experimental rat model of moderate controlled cortical impact (CCI) while considering different time points (1 day, 3 days, 7 days, and 10 days) assessing the acute and sub-acute phase after injury. Our results revealed a new family of lipids, the acylcarnitines, as TBI-lipid related markers, with maximum expression at 3 days after impact and main colocalization within resident microglia of the brain. Furthermore, our experiments highlighted the upregulation of these acylcarnitine lipids, secreted by microglia, in the ipsilateral substantia nigra, the main region in the brain affected in the Parkinson’s disease (PD). Considered as a “Silent Epidemic”, TBI is a leading cause of death and disability with an estimated worldwide incidence of 69 million cases annually1. The biological processes occurring post-injury are divided into two main phases: primary and secondary. The primary phase, is mainly due to the direct mechanical insult to the brain resulting in shearing or stretching of the brain tissue. The secondary phase is characterized by diffuse axonal injury and inflammation2. In neurodegenerative diseases/injuries, lipids are believed to possess a crucial role in the neuroinflammatory process within the secondary phase. In a stroke, for instance, several cytokines, including TNF-alpha, impact phospholipid metabolism and result in the production of ceramides and free radicals, that further injure the brain3. The phenomena of lipid metabolism dysfunction have also been implicated in numerous other neurodegenerative disorders such as Alzheimer’s (AD) and Parkinson (PD) diseases4 and CNS injury, including TBI5.Thus, monitoring the lipid distribution in such injury models is of major importance to better characterize of the impacted area, along with opening new therapeutic windows for any future possible treatment. For such monitoring, Matrix Assisted Laser Desorption/Ionization-mass spectrometry imaging (MALDI-MSI) has allowed detecting the spatial distribution of a broad panel of lipids within a tissue sample in one single experiment6. In a rat ischemic/reperfusion injury model, an increase in ceramide lipid (d18:1/18:0) was detected when compared to sham non-injured condition7. Sparvero et al. monitored the spatial distribution of cardiolipins species within the brain regions in a rat model of CCI8. Several of the cardiolipin species, including CL(72:6) and CL(74:7), showed a decrease in expression within the injured hemisphere of the brain when compared to the contralateral non-injured one in the hippocampal and thalamic regions. In similar decreasing expression pattern, several phosphatidylinositol family

members, especially PI(38:5) and PI(38:4) were identified8. Roux et al., conducted a temporal study (1 vs 7 days) to visualize the differences in lipid expression up to 7 days after CCI via MALDI MSI evaluating abundant lipid mass range from 630-950 Da in the positive ion mode9. In the grey matter, the most abundant lipid changes visualized included ceramide, diacylglycerol, and cholesteryl ester species. Ceramides showed the highest expression at 1-day post-injury and then this expression decreases gradually to reach the lowest levels at 7 days post-injury. Diacylglycerol showed minimum expression at 1 day, reached the highest expression at 3 days, and then decreased again at 7 days post-injury. In a blast brain injury model, ganglioside GM2 showed a major overexpression within the hippocampus, thalamus, and hypothalamus of mice after a single mild blast exposure10. Thus, from one single blast exposure, there was an altered lipid expression within different compartments of the brain. Thus, studying lipid changes in a mild-moderate open head injury model is of major importance. In the present study, we performed a spatiotemporal MALDI MSI analysis on a rat CCI model. The experiments were conducted by considering the acute and subacute phases post injury (1 day, 3 days, 7 days, and 10 days) and considering different regions within the injured brain. Contrary to previously performed MSI studies on TBI, we used a high spectral resolution mass spectrometer instrument assessing both the abundant m/z range (600-900 Da) species as well as the lower mass range (99% purity) palmitoyl-L-carnitine and oleoyl-L-carnitine were purchased from Avanti Polar Lipids (Alabama, USA). OCT embedding matrix was obtained from CellPath (Powys, United Kingdom). Palmitoyl-L-carnitine chloride (>98% purity) was purchased from Santa Cruz Biotechnology (Heidelberg. Germany). Polylysine-coated slides were purchased from Thermo-Scientific (Braunschweig, Germany). Animals. Experiments were conducted on adult male Sprague– Dawley rats (225–250g, 7-8 weeks old) in accordance with the National Institute of Health Guidelines for Animal Research (Guide for the Care and Use of Laboratory Animals) and approved by the Institutional Animal Care and Use Committee (IACUC) at the American University of Beirut (AUB). The rats were maintained and housed under pathogen-free conditions at room temperature in a 12h/12h (light/dark) cycles with 24 hour access to food and water, and each 3 rats were grouped together in one cage. Animals were provided with constant temperature and humidity control at the AUB Animal Care Facility (ACF). All surgical procedures were conducted under deep anesthesia. Experimental Design and Statistical rational. Experiments were performed with biological replicates (n=3). Five cohorts of animals were used including the sham cohort and 4 experimental moderate injury cohorts (1 day, 3 days, 7 days, and 10 days). Controlled Cortical Impact (CCI), Brain Harvesting, and Tissue Preparation. The CCI procedure was performed as previously described11; rats were kept for 1 day, 3 days, 7 days, or 10 days post-injury. At each of the mentioned time points, the rats (n=3) were anesthetized by isoflurane overdose and sacrificed by decapitation. Sham animals received similar surgical preparation, but without any craniotomy of impact. The complete brain was removed and slowly frozen using previously cold iso-pentane on dry-ice; frozen brains were stored at -80 °C until further experimentation. Once needed, the brains were cryosectioned using a cryostat (Leica Microsystems, Nanterre, France) with a tissue thickness of 20 µm. Tissue samples were serially collected and thaw-mounted on Polylysine coated glass slides and stored at -80 °C until further experimentation. MALDI MSI Data Acquisition. 15 min before use, selected tissue sections were placed under vacuum in a desiccator to ensure complete drying. DHB matrix was prepared at a concentration of 20 mg/ml (70:30, v/v) ratio of MeOH/0.1%TFA in H2O. Using a nebulizer which was constructed in-house12, the matrix was manually sprayed on the tissue sections for 12 min, with air pressure=1 bar and flow rate= 300 µl/hr. Images were acquired in the positive ion mode using a MALDI-LTQ-Orbitrap-XL (Thermo Fisher Scientific, Bremen, Germany) operated by Xcalibur software version 2.0.7 (Thermo Fisher Scientific) This system is equipped with a commercial N2 laser (LTB Laser Technik, Berlin, Germany) which operates at λ = 337 nm with a maximum repetition rate

of 60 Hz. The spectral-mass resolution used in our images was 30 000 (centered at m/z 400) in the mass range of 300-1000 Da. Each spectrum was acquired by averaging 10 scans with 1 microscan per step with a raster step size (lateral resolution) of 70 μm. The instrument was calibrated externally in the normal mass range (m/z 150-2000) using a ProteoMass MALDI calibration kit. MALDI MSI Data Processing. All images were converted to vendor-neutral *.imzML format using the plug-in available in ImageQuest software version 1.1.0 build S4 (Thermo Fisher Scientific). SCiLS lab MVS software, version 2018b Premium 3D release 6.01.10194 (Bruker Daltonics, Bremen, Germany) was then used to process the uploaded vendor-neutral files. In this SCiLS Lab version, there is no baseline removal option for vendor-neutral format, and the peaks are automatically imported. Total ion count (TIC) method was used for the normalization and m/z intervals were automatically set at ± 25 mDa. All spectra were aligned to the Move-Peaks-to-LocalMax feature with the reference sample dataset chosen arbitrarily, and the resulting peak list was used in all the subsequent analyses. All images were plotted with weak denoising and intensity scales in figures represent the expression intensity of the plotted m/z across the tissue sample. Spatial segmentation was performed using the bisecting kmeans algorithm with Manhattan distance metric13. Based on the obtained spatial segmentation, we defined our regions to perform the principal component analysis (PCA). The same regions were used to perform colocalization analysis which was carried out using Pearson’s correlation analysis, and Receiveroperating-characteristic (ROC) analysis as previously described11. MS/MS lipid Identification of acylcarnitine lipids. Four consecutive 20-µm tissue sections containing the injured cortical area were obtained using a cryostat (as previously described) and collected in an eppendorf tube for lipid extraction performed using Folch method. 60 µl of CHCl3, 30 µl of MeOH, and 30 µl of H2O were added to the four tissue sections in the mentioned order. The tubes samples were then mixed well by vortex for 1 min, followed by 5 min sonication, and finally centrifuged at 10 000 rcf for 10 min. After centrifugation, the lower phase containing the CHCl3 extracts were collected in a new tube. The upper phase was subjected a second time to the same mentioned CHCl3 extraction process, in order to ensure the maximum extraction possible. Palmitoylcarnitine and oleoylcarnitine standards obtained from Avanti Polar Lipids were prepared at a concentration of 50 µg/µL in CHCl3:MeOH (50:50, v/v). MSn of lipid extracts and lipid standards were performed using MALDI-LTQ-OrbitrapXL in the positive ion mode. Briefly, 1 µl of lipid extract were pipetted onto the MALDI target followed by 1 µl of DHB matrix (prepared as previously mentioned) and directly mixed on the target. Full scans were then acquired at a spectral resolution of 100 000 (centered at m/z 400, similar as previously mentioned) and in the same mass range used to acquire the ion images (m/z 300-1000) by averaging 10 scans acquired with 2 microscans per step. In a targeted approach, the precursor ion was isolated using an isolation window of ±1 and 2 Da and fragments were scanned with a maximum accumulation time between 120 and 200 ms. For the remaining lipid species, preliminary assignments were performed based on accurate mass measurements by interrogation with the METLIN metabolomics database (public domain: http:// metlin.scripps.edu/) using a threshold of ±4 ppm. The molecular

ACS Paragon Plus Environment

Page 2 of 9

Page 3 of 9 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

Analytical Chemistry structure of the acylcarnitines found in the figures of the manuscript were drawn using Chemdraw Professional v16.0. Immunofluorescence Experiments. Immunofluorescence experiments were performed on brain slices directly consecutive to the lesion section MALDI image at 3 days postinjury, only 20 µm away. These staining aimed to differentiate between resident microglia and bone marrow-derived macrophages by staining for Mac-2 and CX3CR1. First, a saturation phase was applied for 1 hour by using blocking buffer constituted of: 1% bovine serum albumin, 1% ovalbumin, 1% normal donkey serum, 0.2% of triton 100X, and 0.1 M glycine all dissolved in PBS 1X. After saturation, primary polyclonal antibodies rabbit Anti-CX3CR1 (1:40 Santa Cruz biotechnology, CliniSciences, Nanterre, France) and mouse Anti Mac-2 (1:40, US Biological lab science, bought from Euromedex, Souffelweyersheim, France) were diluted in the same blocking buffer and added to the same tissue section in order to perform the double staining. In addition, these were not added to the negative control where only the blocking buffer was applied. The primary antibodies were then kept incubated at 4°C overnight. The following day, the sections were washed 3 times with PBS 1x. Secondary antibodies Alexa Fluoro donkey anti-rabbit and Alexa Fluoro donkey anti-mouse (Life Technologies, ThermoFisher Scientific, Courtaboeuf, France) were then added and incubated for 1 hour at 37°C. The secondary antibodies were both prepared in the same buffer as previously mentioned blocking buffer but with no 0.1M glycine, nor 0.2% triton 100X. After 10 min, several washes using PBS 1X were performed and Hoechst solution (1:10000) was added. After, Soudan black 0.3% was added for 10 min to decrease the background signal. The slide was then washed 1 more time with PBS 1X and Dako fluorescent mounting

medium was added before adding coverslip. All images were then captured using confocal microscopy. RESULTS MSI followed by spatial segmentation reveals lipid lesion-specific expression. Tissue sections from rostral, lesion, and caudal positions of non-injured rat brain (sham) and open head injured rat brain at 1 day, 3 days, 7 days, and 10 days post injury were subjected to lipid MALDI-MSI using a MALDILTQ-Orbitrap-XL. All rat replicates received an injury in the same location on the right parietal cortex using an impactor tip of 2 mm surface area. As a result, we were able to image lipids on the same rat brain atlas14 position (Bregma -3.60 mm) corresponding to the lesion section. Also, position (bregma 1.80 mm) and position (bregma -5.60 mm) were chosen for the rostral and caudal sections, respectively. Due to the impactor tip size, the selected rostral and caudal tissue sections had little or no tissue loss when compared to the lesion section. After MALDI-MSI, unsupervised spatial segmentation was performed using Manhattan distance calculation method to visualize regions sharing similar lipid distribution within the same tissue, and in comparison, with the remaining time points and tissue positions (Figure 1A). By dissecting further, the cluster tree, a cluster emerged corresponding to the injured cortical tissue (cluster 6742-yellow) can be observed (Figure 1B). This injury-unique cluster is localized in the microscopically observed injured area and had little to no presence in all three sections (rostral, lesion, and caudal) of the sham brain as seen in the optical images (Figure 1C). In addition, cluster 4396 (light blue) also shows distinct localization within the injury site.

Figure 1: MALDI MSI followed by Unsupervised Spatial Segmentation Reveals Lipid Lesion-specific Expression. A) Rat brain atlas images selected for the rostral, lesion and caudal sections of the study. Spatially segmented cluster images after processing lipid MALDI MSI data of rostral, lesion, and caudal sections from sham brain and 1 day, 3 day, 7 day, and 10 day injured brains. B) Cluster tree corresponding to the imaged sections based on Manhattan distance calculation method. C) Optical scans of the lesion section used in figure 1A for all conditions, along with the co-registered optical and spatially segmented images. However, it is present not only in the cortical tissue, but also in cavities formed after injury. Inter-comparison between the time points shows a wider spread of this injury-related cluster at 3 days post-injury, as the cluster can now be also observed in the rostral and caudal sections. Co-registration of optical

images of the MALDI imaged brains with the spatial segmentation results of the lesion section at all time points shows the precise co-localization between the histologically altered injury site and the obtained injury-related clusters (Figure 1C).

ACS Paragon Plus Environment

Analytical Chemistry 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

To further validate our obtained results, the same unsupervised spatial segmentation approach was applied only on lesion sections from all three replicates in one batch of analysis (Figure S1A). Injured cortical tissue at all time points and in all three replicates grouped in the same cluster (cluster 8087-green) with very little or no distribution in the sham. To visualize the different and similar profiles corresponding to the area of injury when compared to the uninjured cortical tissue, PCA was applied on manually drawn cortical regions between all conditions of all lesion sections in all three replicates (Figure S1B). The data corresponding to the 3 selected cortical

tissue region from sham clearly separate from the remaining clusters corresponding to the injured cortical tissue (Figure S1C). Although this result shows a clear difference in lipid content between sham and injured tissue, no clear separation was observed between different time points post injury. The PCA corresponding to all injured tissues at different time points tend to mix thus yielding an unclear observation for the lipid profile regarding injury at different time points.

Figure 2: MALDI MSI reveals m/z ions overexpressed in injured tissue and not in sham. Plotted Ion images of m/z 780.56 and 666.48 along with their corresponding intensity box plot expression for all three lesion sections replicates in all conditions: sham, 1 day, 3 days, 7 days, and 10 days. lesion-specific signatures and visualize their variation in time. MALDI MSI shows altered lipid expression between With a threshold of 0.75, the ROC analysis yielded 82 peaks injury and non-injury. PCA analysis allowed a separation allowing to distinguish injured tissue from non-injured one between injured cortical tissue and non-injured sham but did not (Table S2). In the most abundant mass range of lipids, which is separate between the selected time points with respect to the between 600-900 m/z, molecules such as 666.48 and 780 .56 ± presence or not of certain lipids, and the variation in expression 25 mDa assigned as PS and PC(36:5) (phosphatidylcholine) level between different time points post-injury. For this reason, respectively had elevated levels within the injury area when receiver operating characteristic (ROC) analysis was performed compared to sham (Figure 2). These two m/z values show a by comparing the same selected regions that were previously slight increase in expression intensity at 1-day post injury. The used in the PCA analysis (Figure S1B) in order to depict unique expression continues to increase until reaching a maximum signals corresponding to both injured tissue and non-injured level at 3 days post injury and followed by a decrease at 7 and tissue (sham). In addition, we aimed to monitor the expression 10 days but still higher than 1-day post injury. This is clearly variation of the injury-related signal in time course along all the visualized by the scatter box plot found in the same figure. replicates of our lesion sections. With a threshold of 0.75, ROC Meanwhile, in the lower mass range, m/z < 600, we identified analysis reveals 67 m/z peaks specific to sham condition when several injury-specific m/z including: 400.35, 459.25, 475.22, analyzed against all four injury time points (Table S1). Of these 487.28, 518.32, 534.30, 544.34, 546.36, and 562.34 ± 25 mDa. 67 ion intervals, two examples m/z 769.57 and 798.53 ± 25 mDa M/z 400.35 was of high interest, as this lipid showed not only corresponding to PA(39:0) (phosphatidic acid) and PS (37:4) an increased and localized expression within the injured cortical (phosphatidylserine) reveal a decreased expression within the tissue of all time points, but also within the corpus callosum of injury site of all time-points when compared to the sham nonboth hemispheres, with a higher intensity in the ipsilateral injured and the non-injured cortical tissue of the opposite hemisphere. m/z 400.35 had no expression in the non-injured hemisphere of the same lesion sections (Figure S2). The box cortical tissue of the contralateral hemisphere within the same scatter plot demonstrates this decrease in expression in injured brain section, neither in the sham. A slight increase within the tissue for all replicates of the study. We then performed the injured tissue is detected at 1 day post-injury, but the maximum same ROC analysis in a similar manner to the previous expression intensity is visualized at 3 days post-injury (Figure conducted one, but now comparing the injury site from all 3). Interestingly, and in in all injury time points, the m/z 400.35 lesion sections at 1 day, 3 days, 7 days, and 10 days combined (green) seems to be outlining the injured tissue along with slight versus the non-injured cortical sham tissue, in order to obtain

ACS Paragon Plus Environment

Page 4 of 9

Page 5 of 9 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

Analytical Chemistry diffusion into the injury core, unlike m/z 518.32 (red), assigned as Lysophosphatidylcholine (16:0) (lysoPC), which shows clear involvement and direct localization within the center of the lesion and not on the injury border. At 7 and 10 days postinjury, and although the expression of m/z 518.32 stays localized within the injury, the intensity expression of m/z 400.32 is no longer localized much within the cortical tissue, rather its expression remains focused in the underlying corpus callosum, but still with increased expression when compared to the sham. Neither m/z 400.35 nor m/z 518.32 were detected in sham sections, thus further confirming their injury related expression.

m/z 400.35 using the sum of spectra from the injured areas. At a threshold of 0.3, 15 peaks were found co-localizing with m/z 400.35 including: 525.39, 518.27, 426.36, 398.32, and 372.31 (Table S3). Interestingly, by preliminary annotation using Metlin online database and later on by MS2, we assigned the peaks at m/z 372, 398, and 426 as tetradecanoylcarnitine, Opalmitoylcarnitine, and oleoylcarnitine respectively which all belong to the same family of lipids as palmitoylcarnitine, the acylcarnitines (Figure S3).

Figure 3: m/z 400 corresponds to Palmitoylcarnitine and is located surrounding the injury core. Co-plotted Ion images of m/z 400.35, 518.32, 866.65, and 769.57 for all three lesion section replicates in all conditions: sham, 1 day, 3 days, 7 days, and 10 days. Injury-related Lipid Biomarker at m/z 400 is Palmitoylcarnitine. We then focused our attention on m/z 400.35 since it has never been described or discussed in the context of TBI in the literature to date. In addition, our results suggest that the expression level of this lipid is relevantly high and only detected post injury with the highest expression at 3 days post injury. We, therefore, decided to orientate our study towards identifying m/z 400.35 along with understanding its spatial distribution along the different time points. To confirm the identity of m/z 400.35, we performed MSn analysis on lipid extracts from the injury sections at 3 days post-impact using MALDI-L TQ-Orbitrap-XL. We assigned m/z 400.35 as Palmitoylcarnitine (C23H46NO4) and its fragmentation pattern is presented in Figure 4A. To further validate the identity of the assigned m/z 400.35, we subjected a commercial palmitoylcarnitine standard to MSn in similar manner as we previously performed on MALDI-LTQ-Orbitrap-XL (Figure 4B) for m/z 400.34 from the lipid extracted at 3 days post-injury. In addition, a search in literature for the MSn fragmentation pattern of m/z 400.35 further confirmed our assignment, as the peaks found by Hulme et al. 15 and those obtained from MSn of palmitoylcarnitine standard were identical to the MSn fragments obtained. The MS2 spectrum of m/z 400.34 yielded one strong intensity peak at m/z 341.268 and one small intensity peak at m/z 239.236. When compared to the molecular structure of palmitoylcarnitine, the peak at 341.268 corresponded to the palmitoylcarnitine with the neutral loss of the Ntrimethylaminegroup (a loss of 59). The second observed signal at m/z 239.236 corresponded to the loss of the glycerol backbone after subjecting the sample to MS3. After obtaining the identity of m/z 400.35 as palmitoylcarnitine, we then aimed to detect other molecules with similar spatial distribution as palmitoylcarnitine within the lesion sections of all time points. To achieve this, we performed co-localization analysis with the

Figure 4: m/z 400.35 is assigned as Palmitoylcarnitine. A) MS/MS spectra showing the fragmentation pattern of m/z 400.35 from lipid extracted from lesion section at 3 days post injury. B) MS/MS spectra showing the fragmentation pattern of m/z 400.35 lipid standard. The difference between palmitoylcarnitine and opalmitoylcarnitine is the presence of a double bond in the carbon side chain of the latter. The difference between palmitoylcarnitine and tetradecanoylcarnitine is the length of the carbon side chain as palmitoylcarnitine is 16:0, tetradecanoylcarnitine and 14:0. Finally, the difference between palmitoylcarnitine (16:0) and oleoylcarnitine (18:1) is the length of the carbon side chain and the presence of double bond in the latter. All three detected acylcarnitine members show the same spatial distribution as palmitoylcarnitine regarding both times point and distribution within the tissue structures (Figure S4). Expression of palmitoylcarnitine in the substantia nigra three days post injury. To detect the presence of palmitoylcarnitine (16:0) in other possible affected regions within the remaining regions of the brain, other than the injured cortical tissue, we plotted the ion images of m/z 400.35 in the rostral and caudal sections of all time points used in this study. Surprisingly, in the caudal section at 3 days post injury and partially within the midbrain area, palmitoylcarnitine intensity was highly elevated in the substantia nigra of the injured

ACS Paragon Plus Environment

Analytical Chemistry 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

hemisphere (Figure 5). The intensity of expression is equivalent to that present within the injured cortical tissue and not detected in the remaining areas of the brain sections at 3

days. In addition, there was an expression of palmitoylcarnitine within the substantia nigra of the caudal sections corresponding to the remaining time points, neither of that found in the sham.

Figure 5: Expression of Palmitoylcarnitine in the substantia nigra three days post injury. Plotted ion images of m/z 400.35 rostral, lesion, and caudal sections in all conditions: sham, 1 day, 3 days, 7 days, and 10 days. The substantia nigra containing the expression of palmityocarnitine is pointed with an orange arrow. Taken together, this finding reveals the unique temporal and spatial expression of palmitoylcarnitine in the substantia nigra at 3 days following the injury. Localization of palmitoylcarnitine with resident microglia of lesion margins. Due to the blood brain barrier breach occurring post injury, circulating macrophages can infiltrate into the injured area of the brain. Besides, resident microglia have been known to travel towards the injury site and form a surrounding barrier. Furthermore, as previously stated, the palmitoylcarnitine expression surrounds the injury core, mainly at 3 days. In order to confirm the localization and correlate it with the resident microglia or infiltrating microglia/macrophages, immunofluorescence staining experiments where performed. For this, we double-stained lesion section at 3 days post-injury with CX3C chemokine receptor 1 (CX3CR1) and beta-galacto macrophages/microglia. However, it is the intensity ratio difference of expression between these two markers which allows the differentiation between them. In fact, cells expressing the high intensity of CX3CR1 when compared to Mac-2 (CXCR1+++/Mac-2+/-) corresponds to resident microglia. On the other hand, the more intense signal of Mac-2 when compared to CX3CR1 correspond to bone marrow-derived macrophages/microglia, thus infiltrating cells into the injured environment 16. Staining of the tissue at 3 days post-injury reveals more intense signal for CX3CR1 on the surrounding of the injury showing that resident microglia have a possible role in the sequestration and isolation of the damaged tissue while the core of the injury shows a more intense signal of Mac-2 thus resembling the infiltrating macrophages (Figure 6). By performing MALDI imaging on consecutive tissue of the one subjected to IF, palmitoylcarnitine clearly co-localizes with the resident microglia of the brain in the same expression pattern, thus forming a barrier around the injured tissue. DISCUSSION Long chain acylcarnitines (LC ACs, including palmitoylcarnitine and oleoylcarnitine) seem to be injuryrelated lipid biomarkers. In fact, during -oxidation, these

molecules are intermediate compounds in the shuttle of free fatty acids into the inner mitochondrial membrane17. Thus, any mitochondrial damage or dysfunction can lead to the accumulation of these acylcarnitines. These LC-ACs have been studied by several groups and in different fields of research. Using MALDI-MSI, palmitoylcarnitine was detected with high abundance at the loci of Salmonella typhimurium infection and mesenteric lymph node (MLN) disruption15.

Figure 6: Localization of Palmitoylcarnitine with resident microglia along the lesion margin. A) Confocal immunofluorescence images of lesion section at 3 days post injury along with the zoom images of the lesion borders stained with two markers: CX3CR1 (green) and Mac-2 (red). B) Superimposed optical image and ion image showing the following m/z ions: 518.32 (green), 400.35 (red), and 866.65 (blue). By culturing MLN cells (including B and T cells, along with macrophages and dendrites) ex-vivo, stimulation with palmitoylcarnitine induced the apoptosis of one subset of Tcells (CD4+CD25+) via the caspase-3/7 pathway. Also, palmitoylcarnitine co-localizes with disrupted immune cells,

ACS Paragon Plus Environment

Page 6 of 9

Page 7 of 9 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

Analytical Chemistry suggesting its possible mode of apoptotic action within the MLN. In MDA-MB-231 breast cancer cell lines, MALDI-MSI revealed the localization of acylcarnitines, mainly palmitoylcarnitine and stearoylcarnitine at m/z 400.3 and 428.3 respectively, along with LysoPC and other lipid species. The two acylcarnitine species were mainly located in the hypoxic regions of the tumor18. In addition, LC ACs are believed to have a pivotal role in the development of atherosclerosis. Treatment of apoE-/- mice with methyl-γ-butyrobetaine (methyl-GBB) phosphate, an inhibitor of L-carnitine biosynthesis and transport, decreased the expression level of short chain (C2 to C4) acylcarnitine by 17 fold, and the expression level of medium chain (C6 to C12) and long chain (C14 to C18) acylcarnitine by 7 fold in aortic tissue. This treatment led to a significant attenuation of atherosclerotic lesions in the whole aorta thus suggesting a possible role of acylcarnitines in the development of atherosclerosis19. Methyl-GBB treatment also decreased the level of infiltration of macrophages and monocytes, thus explaining a possible mode of action for acylcarnitines by activating immune cells in the case of atherosclerosis. In the brain, acylcarnitines are major contributors in processes such as fatty acid metabolism and ketosis, thus resulting in significant impact on brain metabolism20. In particular, palmitoylcarnitine is involved in synthesis of several lipids which in turn aid neural membranes and brain plasticity. In addition, palmitoylcarnitine is known to have an inhibitory effect on protein kinase C (pkc), a major signaling protein involved in cell growth21. Thus, there exists a central question of the lipid role in the physiological mechanisms of injury in the CNS, and its previously shown that they act as central players in the regeneration process22. To address this question in our presented study, we applied a lipidomic approach using MALDI MSI. Our study also revealed the presence of long and medium chain acylcarnitines, along with LysoPCs that showed overexpression only in the injured conditions and not in the non-injured conditions. LysoPC molecules possess a “find-me” and “eat-me” bivalent function for the phagocytosis of apoptotic cells23–25. These molecules attract phagocytes to the area apoptotic cells, and at the same time, act as opsonins, to tag dying cells to be phagocytosed by infiltrating cells. Using MSI to detect the presence of lysoPC, coupled with IF, these previous findings clearly explain the colocalization between infiltrating macrophages and lysoPC as the latter is mediating the phagocytic effect of macrophages within the injury core to enter and clear apoptotic cells along with injury debris. On the other hand, resident microglia have been demonstrated to rush to the lesion and form a cellular barrier that will sequester the injury site and prevent the exacerbation of the damage26. In TBI, we have shown that palmitoylcarnitine also has an injuryrelated distribution, but not within the core, but rather along the injury borders. This expression was colocalized with the resident microglia that migrate to surround the injury core. Thus, we believe, that acylcarnitine mapping reveals a wellsegmented immune distribution, as they are mainly colocalizing with resident microglia, and not with the infiltrating macrophages. This accumulation of acylcarnitine, may show a dual role for these molecules. On one side, they may aid the proinflammatory response that occurs within the injury site, as they have been previously described be implicated in inflammation. On the other hand, they may have a crucial role, in isolating the injured cells from the non-injured ones, and thus result in a decrease of damage spread. We have previously reported the involvement acylcarnitine in spinal cord injury (SCI)27, and surprisingly, acylcarnitines have shown to follow the same

spatial distribution and expression pattern in both TBI and SCI injury models, by enclosing the injury core and colocalizing with resident microglia. Taken altogether, these findings show the common lipid response that is in common between both types of CNS injury forms. Finally, a better understanding of the molecular mechanisms occurring in brain regions after remote injury is of high interest to develop innovative therapeutic strategies for the long term consequences of TBI. Notably, it has been established that TBI is a risk factor of PD28. TBI is associated with a 44% increased risk of developing PD over 5 to 7 years29. We demonstrated that expression levels of palmityolcarnitine within the ipsilateral substantia nigra were similar to that found in the injury site. This expression could possibly hint to a pathophysiological link between TBI and PD, as the latter is mainly characterized by dopaminergic neuronal death30. This is the first study to shed light on the possible link between TBI and PD at such an early time point in the acute phase of TBI. Indeed, these preliminary findings suggest the involvement of palmityolcarnitine in the early inflammation phase of the microglia but also more broadly the involvement of the β-oxidation and its carnitine shuttle through the mitochondria for energy production from fatty acid as a key mechanism to avoid programmed cell death of dopaminergic neurons leading to PD. In conclusion, our results shed light on the importance of using high throughput studies such as lipid MALDI MSI to visually detect the spatial distribution of injury-related lipids and their possible link to key immune players post-impact to the CNS. We are the first to characterize the involvement of palmitoylcarnitine, and several other acylcarnitine family members, within the context of TBI and have shown its spatial and temporal distribution with resident microglia surrounding the injured cortical tissue. These preliminary findings may have potential in understanding how people who have sustained a TBI during their lifetime, develop PD in later stages of life. ASSOCIATED CONTENT Supporting Information The Supporting Information is available free of charge on the ACS Publications website at DOI: Supp_tables_figures (PDF) AUTHOR INFORMATION * Corresponding authors [email protected] and [email protected] ORCID Michel Salzet: 0000-0003-4318-0817 ; Isabelle Fournier: 00000003-1096-5044 Author Contributions Conceptualization: F.K., , I.F. and M.S.; Methodology, I.F. and M.S.; Software, J.Q. and K.M.; Validation, A.R-R., I.F., J.Q., K. M. and M.S.; Formal Analysis, J.Q. and K.M.; Investigation, A.R-R., J.Q., K.M. and S.A. ; Data curation, I.F., J.Q., K.M. and M.S.; Writing - Review & Editing, D.D., F.K., I.F., J.Q., K.M., and M.S. ; Supervision, I.F. and M.S.; Project Administration, I.F. and M.S.; Funding Acquisition, I.F., K.Z. and M.S. Notes The authors declare no competing financial interest. DATA AND SOFTWARE AVAILABILITY All MALDI imaging datasets were uploaded to ProteomeXchange Consortium via the PRIDE database and was assigned the dataset identifier PXD011262.

ACS Paragon Plus Environment

Analytical Chemistry 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

ACKNOWLEDGMENT This research was supported by funding from Ministère de lʼEnseignement Supérieur, de laRecherche et de lʼInnovation (MESRI), Institut National de la Santé et de la Recherche Médicale (Inserm) and Université de Lille. KM acknowledges funding from the Association of Scientific Orientation and Specialization (ASOS Scholarship). REFERENCES (1) Dewan, M. C.; Rattani, A.; Gupta, S.; Baticulon, R. E.; Hung, Y.-C.; Punchak, M.; Agrawal, A.; Adeleye, A. O.; Shrime, M. G.; Rubiano, A. M.; et al. Estimating the Global Incidence of Traumatic Brain Injury. J. Neurosurg. 2019, 130 (4), 1080– 1097. https://doi.org/10.3171/2017.10.JNS 17352. (2) Das, M.; Mohapatra, S.; Mohapatra, S. S. New Perspectives on Central and Peripheral Immune Responses to Acute Traumatic Brain Injury. J. Neuroinflammation 2012, No. c, 1–12. (3) Adibhatla, R. M.; Dempsy, R.; Hatcher, J. F. Integration of Cytokine Biology and Lipid Metabolism in Stroke. Front. Biosci. 2008, 13, 1250–1270. (4) Wenk, M. R. Erratum: The Emerging Field of Lipidomics. Nat. Rev. Drug Discov. 2005, 4 (7), 594–610. https://doi.org/10.1038/nrd1776. (5) Adibhatla, R. M.; Hatcher, J. F. Altered Lipid Metabolism in Brain Injury and Disorders. In Lipids in Health and Disease; Springer Netherlands: Dordrecht, 2008; Vol. 49, pp 241–268. https://doi.org/10.1007/978-1-4020-8831-5_9. (6) Zemski Berry, K. A.; Hankin, J. A.; Barkley, R. M.; Spraggins, J. M.; Caprioli, R. M.; Murphy, R. C. MALDI Imaging of Lipid Biochemistry in Tissues by Mass Spectrometry. Chem. Rev. 2011, 111 (10), 6491–6512. https://doi.org/10.1021/cr200280p. (7) Hankin, J. A.; Farias, S. E.; Barkley, R. M.; Heidenreich, K.; Frey, L. C.; Hamazaki, K.; Kim, H.-Y.; Murphy, R. C. MALDI Mass Spectrometric Imaging of Lipids in Rat Brain Injury Models. J. Am. Soc. Mass Spectrom. 2011, 22 (6), 1014–1021. https://doi.org/10.1007/s13361-011-0122-z. (8) Sparvero, L. J.; Amoscato, A. A.; Fink, A. B.; Anthonymuthu, T.; New, L. A.; Kochanek, P. M.; Watkins, S.; Kagan, V. E.; Bayır, H. Imaging Mass Spectrometry Reveals Loss of Polyunsaturated Cardiolipins in the Cortical Contusion, Hippocampus, and Thalamus after Traumatic Brain Injury. J. Neurochem. 2016, 139 (4), 659–675. https://doi.org/10.1111/jnc.13840. (9) Roux, A.; Muller, L.; Jackson, S. N.; Post, J.; Baldwin, K.; Hoffer, B.; Balaban, C. D.; Barbacci, D.; Schultz, J. A.; Gouty, S.; et al. Mass Spectrometry Imaging of Rat Brain Lipid Profile Changes over Time Following Traumatic Brain Injury. J. Neurosci. Methods 2016, 272, 19–32. https://doi.org/10.1016/j.jneumeth.2016.02.004. (10) Woods, A. S.; Colsch, B.; Jackson, S. N.; Post, J.; Baldwin, K.; Roux, A.; Hoffer, B.; Cox, B. M.; Hoffer, M.; Rubovitch, V.; et al. Gangliosides and Ceramides Change in a Mouse Model of Blast Induced Traumatic Brain Injury. ACS Chem. Neurosci. 2013, 4 (4), 594–600. https://doi.org/10.1021/cn300216h. (11) Mallah, K.; Quanico, J.; Trede, D.; Kobeissy, F.; Zibara, K.; Salzet, M.; Fournier, I. Lipid Changes Associated with Traumatic Brain Injury Revealed by 3D MALDI-MSI. Anal. Chem. 2018, 90 (17), 10568–10576. https://doi.org/10.1021/acs.analchem.8b02682. (12) Franck, J.; Quanico, J.; Wisztorski, M.; Day, R.; Salzet, M.; Fournier, I. Quantification-Based Mass Spectrometry Imaging of Proteins by Parafilm Assisted Microdissection. Anal. Chem. 2013, 85 (17), 8127–8134. https://doi.org/10.1021/ac4009397. (13) Alexandrov, T.; Kobarg, J. H. Efficient Spatial Segmentation of Large Imaging Mass Spectrometry Datasets with Spatially Aware Clustering. Bioinformatics 2011, 27 (13), i230–i238.

https://doi.org/10.1093/bioinf ormatics/btr246. (14) Paxinos, G.; Watson, C. The Rat Brain in Stereotaxic Coordinates. Acad. Press 1998, No. 1, 1–474. https://doi.org/10.1007/s13398-014-0173-7.2. (15) Hulme, H. E.; Meikle, L. M.; Wessel, H.; Strittmatter, N.; Swales, J.; Thomson, C.; Nilsson, A.; Nibbs, R. J. B.; Milling, S.; Andren, P. E.; et al. Mass Spectrometry Imaging Identifies Palmitoylcarnitine as an Immunological Mediator during Salmonella Typhimurium Infection. Sci. Rep. 2017, 7 (1), 2786. https://doi.org/10.1038/s41598-017-03100-5. (16) Wang, X.; Cao, K.; Sun, X.; Chen, Y.; Duan, Z.; Sun, L.; Guo, L.; Bai, P.; Sun, D.; Fan, J.; et al. Macrophages in Spinal Cord Injury: Phenotypic and Functional Change from Exposure to Myelin Debris. Glia 2015, 63 (4), 635–651. https://doi.org/10.1002/glia.22774. (17) Schooneman, M. G.; Vaz, F. M.; Houten, S. M.; Soeters, M. R. Acylcarnitines. Diabetes 2013, 62 (1), 1–8. https://doi.org/10.2337/db12-0466. (18) Chughtai, K.; Jiang, L.; Greenwood, T. R.; Glunde, K.; Heeren, R. M. A. Mass Spectrometry Images Acylcarnitines, Phosphatidylcholines, and Sphingomyelin in MDA-MB-231 Breast Tumor Models. J. Lipid Res. 2013, 54 (2), 333–344. https://doi.org/10.1194/jlr.M027961. (19) Vilskersts, R.; Kuka, J.; Liepinsh, E.; Makrecka-Kuka, M.; Volska, K.; Makarova, E.; Sevostjanovs, E.; Cirule, H.; Grinberga, S.; Dambrova, M. Methyl-γ-Butyrobetaine Decreases Levels of Acylcarnitines and Attenuates the Development of Atherosclerosis. Vascul. Pharmacol. 2015, 72, 101–107. https://doi.org/10.1016/j.vph.2015.05.005. (20) Jones, L. L.; McDonald, D. A.; Borum, P. R. Acylcarnitines: Role in Brain. Prog. Lipid Res. 2010, 49 (1), 61–75. https://doi.org/10.1016/j.plipres.2009.08.004. (21) Nakadate, T.; Blumberg, P. M. Modulation by Palmitoylcarnitine of Protein Kinase C Activation. Cancer Res. 1987, 47 (24 Pt 1), 6537–6542. (22) Knobloch, M.; Braun, S. M. G.; Zurkirchen, L.; von Schoultz, C.; Zamboni, N.; Araúzo-Bravo, M. J.; Kovacs, W. J.; Karalay, Ö.; Suter, U.; Machado, R. A. C.; et al. Metabolic Control of Adult Neural Stem Cell Activity by Fasn-Dependent Lipogenesis. Nature 2012, 493 (7431), 226–230. https://doi.org/10.1038/nature11689. (23) Mueller, R. B.; Sheriff, A.; Gaipl, U. S.; Wesselborg, S.; Lauber, K. Attraction of Phagocytes by Apoptotic Cells Is Mediated by Lysophosphatidylcholine. Autoimmunity 2007, 40 (4), 342–344. https://doi.org/10.1080/0891693070 1356911. (24) Lauber, K.; Bohn, E.; Kröber, S. M.; Xiao, Y.; Blumenthal, S. G.; Lindemann, R. K.; Marini, P.; Wiedig, C.; Zobywalski, A.; Baksh, S.; et al. Apoptotic Cells Induce Migration of Phagocytes via Caspase-3-Mediated Release of a Lipid Attraction Signal. Cell 2003, 113 (6), 717–730. (25) Kim, S. J.; Gershov, D.; Ma, X.; Brot, N.; Elkon, K. B. IPLA(2) Activation during Apoptosis Promotes the Exposure of Membrane Lysophosphatidylcholine Leading to Binding by Natural Immunoglobulin M Antibodies and Complement Activation. J. Exp. Med. 2002, 196 (5), 655–665. (26) Hines, D. J.; Hines, R. M.; Mulligan, S. J.; Macvicar, B. A. Microglia Processes Block the Spread of Damage in the Brain and Require Functional Chloride Channels. Glia 2009, 57 (15), 1610–1618. https://doi.org/10.1002/ glia.20874. (27) Quanico, J.; Hauberg-Lotte, L.; Devaux, S.; Laouby, Z.; Meriaux, C.; Raffo-Romero, A.; Rose, M.; Westerheide, L.; Vehmeyer, J.; Rodet, F.; et al. 3D MALDI Mass Spectrometry Imaging Reveals Specific Localization of Long-Chain Acylcarnitines within a 10-Day Time Window of Spinal Cord Injury. Sci. Rep. 2018, 8 (1), 16083. https://doi.org/10.1038/s41598-018-34518-0. (28) Gardner, R. C.; Byers, A. L.; Barnes, D. E.; Li, Y.; Boscardin, J.; Yaffe, K. Mild TBI and Risk of Parkinson Disease. Neurology 2018, 90 (20), e1771–e1779.

ACS Paragon Plus Environment

Page 8 of 9

Page 9 of 9 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

Analytical Chemistry https://doi.org/10.1212/WNL.0000000000005522. (29) Gardner, R. C.; Burke, J. F.; Nettiksimmons, J.; Goldman, S.; Tanner, C. M.; Yaffe, K. Traumatic Brain Injury in Later Life Increases Risk for Parkinson Disease. Ann. Neurol. 2015, 77

(6), 987–995. https://doi.org/10.1002/ana.24396. (30) Naoi, M.; Maruyama, W. Cell Death of Dopamine Neurons in Aging and Parkinson’s Disease. Mech. Ageing Dev. 1999, 111 (2–3), 175–188.

Table of Contents artwork

ACS Paragon Plus Environment