Mass Spectrometry Imaging and GC-MS Profiling of the Mammalian Peripheral Sensory-Motor Circuit
- 879 Downloads
Matrix-assisted laser desorption/ionization (MALDI) mass spectrometry imaging (MSI) has evolved to become an effective discovery tool in science and clinical diagnostics. Here, chemical imaging approaches are applied to well-defined regions of the mammalian peripheral sensory-motor system, including the dorsal root ganglia (DRG) and adjacent nerves. By combining several MSI approaches, analyte coverage is increased and 195 distinct molecular features are observed. Principal component analysis suggests three chemically different regions within the sensory-motor system, with the DRG and adjacent nerve regions being the most distinct. Investigation of these regions using gas chromatography-mass spectrometry corroborate these findings and reveal important metabolic markers related to the observed differences. The heterogeneity of the structurally, physiologically, and functionally connected regions demonstrates the intricate chemical and spatial regulation of their chemical composition.
KeywordsMass spectrometry imaging Gas chromatography-mass spectrometry Peripheral sensory-motor system Dorsal root ganglia Spinal nerve
The spatial-temporal chemical and structural heterogeneity of the nervous system is enormous, making the evaluation of the tissue or organ metabolome at spatial resolutions approaching the single cell level problematic as the metabolome varies in a region and cell-specific manner. Mass spectrometry imaging (MSI) is a multifaceted analytical approach that allows the spatial localization and characterization of a broad range of analytes at the tissue, cellular, and even subcellular levels [1, 2, 3, 4, 5, 6, 7]. Caprioli has played a pivotal role in the introduction, development, and popularization of matrix-assisted laser desorption/ionization (MALDI)-MSI [4, 5, 8, 9, 10]. The approach typically delivers multiplexed information, with tens to hundreds of compounds detected at each location within a sample. Because of the chemical complexity of tissues, this analyte coverage is not enough to characterize the complete metabolome, peptidome, or proteome. Although a range of additional analyte separation and conditioning steps have been developed, they are not easily implemented because of the thousands of individual spots sampled across a sample surface during MSI.
To address the problem of low analyte coverage, several off-line and on-line strategies have been implemented, including pre-separation of ions using ion mobility approaches [11, 12, 13, 14], high resolution mass analyzers [8, 15, 16, 17, 18], and liquid chromatography pre-separation of analytes . One of the more obvious strategies to increase analyte coverage is to combine several MSI methods, each optimized for specific analyte classes [19, 20, 21, 22]. For example, MALDI-MSI uses a variety of matrices that preferentially enhance the desorption/ionization of specific classes of analytes . At times, the presence of MALDI matrix complicates the detection of metabolites because matrix ions may dominate the mass spectra and/or overlap with endogenous signals. An older approach, laser desorption/ionization (LDI), can be successfully used for some analytes while generating less background ions. As MALDI-MSI and LDI-MSI are performed on the same instrument, combining the two is a straightforward methodology to implement in discovery experiments. Of course, because this combination increases the time required to probe samples, smaller specimens of many brain regions or endocrine structures such as the pituitary [24, 25, 26] and spinal cord [27, 28, 29, 30, 31] are best suited to this strategy.
For this work, we performed MSI and other investigations of several areas of the mammalian sensory-motor system comprising the well-defined and compact dorsal root ganglion (DRG) and adjacent nerves (dorsal root, ventral root, and spinal nerve) in order to determine how the chemistry of different regions correlates to known structural motifs. A variety of animal nerves have been investigated by direct and indirect MS approaches, including the stomatogastric nerve of the blue crab Callinectes sapidus , the major connectives in the mollusk Aplysia californica , the mouse sciatic nerve , the rodent optic nerve , and the rat tibial  and lumbar nerves .
The DRG is a morphologically and biochemically complex structure containing thousands of cells having various functions, including proprioception and nociception. The DRG is a well-established neurobiological model that is widely-used in investigations of the mammalian nervous system’s normal and pathologic activities [36, 38, 39, 40, 41]. There are more than 5000 papers that describe different aspects of DRG morphology, physiology, and biochemistry, including MS-based studies of their elemental , metabolite [36, 43, 44, 45], and protein [46, 47, 48, 49] contents. Axons of most DRG-located neurons are bifurcated, with the resulting branches extending inside the nerves towards the periphery (e.g., skin) and spinal cord. The nerves are complex structures where many neurites are ensheathed by Schwann cells and grouped in bundles surrounded by the epineurium and the perineurium. The presence of high lipid content in the myelin sheath makes these nerves difficult targets for direct MS investigation of endogenous peptides and proteins. However, the important role of nerves in normal and pathologic peripheral information collection and centrally generated signal conduction make them important targets for metabolomic [36, 43, 44] and proteomic  studies. In this work, after optimization of tissue sampling protocols, both LDI-MSI and MALDI-MSI were utilized to investigate the chemical heterogeneity of the functionally connected regions of the mammalian peripheral sensory-motor network.
All reagents were purchased from Sigma-Aldrich (St. Louis, MO, USA) unless otherwise specified.
All procedures related to animal handling and euthanasia were performed in accordance with local, state, and federal regulations and approved by the Illinois Institutional Animal Care and Use Committee. DRGs, spinal and sciatic nerves, and ventral and dorsal roots were surgically dissected from adult 6–12 wk-old Long-Evans/BluGill male and female rats (University of Illinois at Urbana-Champaign), euthanized by decapitation. To reduce the detrimental metabolic effects of a stopped blood flow, 40–60 mL of ice-cold Gey’s balanced salt solution (mGBSS) was injected under the skin above the location of the DRGs of interest (the lumbar area) immediately after decapitation. The spinal canal was surgically opened, the spinal cord removed, and DRGs with adjacent nerves were individually isolated and placed into ice-cold mGBSS containing (in mM): 1.5 CaCl2, 4.9 KCl, 0.2 KH2PO4, 11 MgCl2, 0.3 MgSO4, 138 NaCl, 27.7 NaHCO3, 0.8 Na2HPO4, and 25 HEPES, pH 7.2. In gas chromatography-mass spectrometry (GC-MS) experiments, a mixture of 33% (v/v) glycerol and 67% mGBSS was used to stabilize the chemistry of the studied system after isolation.
MSI was performed on sections of the DRG and adjacent nerves. The DRGs and nerves were frozen and sectioned in 10-μm-thick slices without chemical fixation at –20°C using a Leica CM 3050 S cryostat (Leica Microsystems, Bannockburn, IL, USA). DRG sections were deposited on room temperature (23–25°C) indium tin oxide coated (ITO) glass slides (Bruker Daltonics, Billerica, MA, USA) and dried in a desiccator at 23–25°C for LDI-MSI. Specimens for MALDI-MSI were spray coated with 5 mg/mL 2-(4-hydroxyphenylazo) benzoic acid (HABA) in methanol/water-containing a 0.1% formic acid and 0.01% trifluoroacetic acid (TFA) (20/80 v/v) mixture or with 10 mg/mL 9-aminoacridine (9-AA) in a water/ethanol/TFA (30/70/1, v/v/v) solution . An artist’s airbrush attached to an in-house nitrogen line was used to apply MALDI matrix solutions onto tissue slices. Depending on the composition of the MALDI matrix, the applications were performed from a 50 to 100 cm distance in several iterations, allowing formation of a thin layer of liquid on top of the slides and their complete drying. Application of MALDI matrix solution was stopped when a thin uniform layer of matrix formed on top of the tissue sections. The quality of coating was visually observed using a Leica MZ 7.5 high-performance stereomicroscope with a 7.9:1 zoom (Leica Microsystems Inc.).
MSI was performed with an Ultraflex II TOF-TOF mass spectrometer (Bruker Daltonics) operating in reflector mode at positive or negative polarities. The laser beam was set to “small”, corresponding to a 2020–4040 μm diameter footprint. The acquired chemical images have either a 25 or 50 μm spatial resolution. Mass spectra were acquired over the m/z 20–2000 mass range. Mass spectrometer calibration was performed using a Peptide Calibration Standard Kit II (angiotensin II, angiotensin I, Substance P, bombesin, ACTH clip 1-17, ACTH clip 18-39, somatostatin 28, bradykinin fragment 1-7, renin substrate tetradecapeptide porcine) obtained from Bruker Daltonics. The standards were deposited onto ITO slide locations nearby the DRG sections. The peptide ions, and in some cases known endogenous ions, such as several of the most common lipids as well as MALDI-matrix ions, were used in the post-processing step for data recalibration. The following were typical performance parameters for representative spots: mass resolution was typically between 1000 and 2500, and the mass error, which depended on location and topology of the sample, ranged between 30 and 120 ppm.
Integrated software was used for MSI data acquisition and processing: FlexControl 3 – control of the instrument and mass spectra acquisition; FlexImaging 3 and 4 – chemical image acquisition, visualization, and data processing; FlexAnalysis 3 – visualization and processing of individual mass spectra as well as batch data processing; and ClinProTools 3 – statistical analysis, including multivariate analysis of acquired data (all programs are from Bruker Daltonics).
L4 and L5 DRGs and adjacent nerves, as well as the distal portion of the sciatic nerve, were isolated from three animals during an ~1 h dissection procedure done on ice. The individual structures were placed in 200 μL of ice cold 33% glycerol, 67% mGBSS solution, and stored on ice. The DRGs and nerves were separated and individually weighed after Kimwipe-assisted removal of excess solution surrounding the tissues. The wet mass of the glycerinated tissues varied between less than 20 (DRGs) to several hundred (distal part of sciatic nerve) milligrams. Metabolite extraction was performed using a modified two-phase approach described in . Briefly, 200 μL of a solution containing 2:1 (v/v) reagent grade methanol and chloroform, cooled to 4°C, was added to each tube. Tissues were homogenized on ice. Homogenates were centrifuged at 18,000 g for 10 min at 4°C. Supernatant was collected and moved to fresh tubes and stored on ice. Chloroform (100 μL) was added to the remaining material (pellet). The pellets were homogenized on ice and 100 μL of water was added. Samples were briefly vortexed and centrifuged at 18,000 g for 10 min at 4°C. Two phases were formed in the extract solution after the centrifugation: an upper phase containing methanol and water, and a lower phase containing chloroform. A glass syringe with a long fine needle (Hamilton, Reno, NV, USA) was used to collect the phases separately and place them into corresponding vials. The lower phase was combined with the supernatant collected during the first extraction and specified as the “L” sample. The upper phase was placed into fresh vials and specified as “T”. All samples were dried in a SpeedVac concentrator (Thermo Scientific, San Jose, CA, USA).
Dried samples were derivatized according to Roessner et al.  with minor modifications: 90 min at 50°C with 80 μL of methoxyamine hydrochloride in pyridine (20 mg mL –1) followed by 60 min treatment at 50°C with 80 μL N-Methyl-N-(trimethylsilyl)trifluoroacetamide (MSTFA). An internal standard (10 μL of C31 fatty acid) was added prior to trimethylsilylation.
A sample volume of 1 μL was injected onto the column with a split ratio of 5:1. The GC-MS system consisted of an Agilent 7890A (Agilent Technologies, Inc., Palo Alto, CA, USA) gas chromatograph, an Agilent 5975C mass selective detector, and an Agilent 7683B autosampler.
Gas chromatography was performed on a 60 m HP-5MS column with a 0.25 mm inner diameter and 0.25 μm film thickness (Agilent) with an injection temperature of 250°C, the interface set to 250°C, and the ion source adjusted to 230°C. The helium carrier gas was set at a constant flow rate of 1.5 ml min–1. The temperature program was 5 min of isothermal heating at 70°C, followed by an oven temperature increase of 5°C min–1 to 310°C, and a final 20 min at 310°C. Mass spectra were recorded in the m/z 50–800 scanning range.
The spectra of chromatogram peaks were compared with the following electron ionization mass spectrum libraries: NIST08 , WILEY08 , and a custom library consisting of 1183 unique mass spectra for common, naturally occurring metabolites (their TMS derivatives, methyl esters, and some compounds that are not amenable to derivatization but also present in biological samples). Authentic standards were used in creation of the custom library. To allow comparison between samples, data were normalized to the internal standard (hentriacontanoic acid) in each chromatogram and fresh weight of each sample. The chromatograms and mass spectra were evaluated using the MSD ChemStation (Agilent) and AMDIS  programs. The retention time and mass spectra were implemented within the AMDIS method formats. Metabolite concentrations are reported as concentrations relative to the internal standard (i.e., target compound peak area divided by peak area of hentriacontanoic acid: Ni = Xi × X–1IS per 100 mg wet weight. The instrument variability was within the standard acceptance limit (5%). Three individual samples collected from three different animals were analyzed using GC-MS. GC-MS was performed in the Roy J. Carver Metabolomics Center at the University of Illinois at Urbana-Champaign.
Partial Least Squares-Discriminant Analysis (OPLS/O2PLS-DA) Model for Evaluation of GC-MS Results
For the metabolite profiling, the instrument variability was within the standard acceptance limit (5%). GC-MS analysis of both polar and nonpolar metabolites detected a total of ~1200 electron ionization mass spectral features in the samples. Of these, 194 compounds were positively identified, with the majority being lipids. Chemometric models were obtained using the SIMCA P+ (12.0) program (Umetrics, Umeå, Sweden). Metabolites with 80% or more missing data were removed from analysis; the rest of the missing values for particular metabolites were imputed with 0. All spurious metabolites derived from column bleed and reagent artifacts were also removed from the data sets.
Due to an uncertain principal component analysis (PCA) model, a supervised PLS analysis was performed on centered data for validation of the generated chemometric models. The number of observations used in chemometric models is typically much lower than the total number of variables; it is possible that the good performance statistics observed in the chemometric models could be due to data overfitting [55, 56, 57]. The obtained PLS-DA model was further analyzed using 7-fold cross-validation ANOVA and response permutation with 900 random reclassifications (random assignment of class labels in order to test whether differences found between groups are significant). The observed PLS-DA model had R2Y = 16.9%, Q2 = –15.3% (negative prediction ability), and variables explained 81.1% (R2X) of total variation [55, 56, 57, 58, 59, 60, 61].
Results and Discussion
DRG Sample Preparation for MSI Analyses
MALDI-MSI of the Mammalian Sensory-Motor Network
As a discovery tool, MALDI-MSI effectively highlights the putative differences in the chemical composition of different biological structures, and this guides further investigations. For example, the image of the distribution of the analyte with m/z 883 demonstrates the biochemical differences between ventral and dorsal root neurons and/or glia-related structures, and allows us to follow its contribution to the formation of the spinal nerve (Supplementary Figure S1a; ventral root component outlined by dashed line). Additionally, visual observation of the distribution of molecular signals helps to uncover potentially interesting low-intensity signals from endogenous compounds (Supplementary Figure S1b). Such signals have a low signal to noise ratio and, typically, are excluded from peak peaking by automatic mass spectra processing.
GC-MS Analysis of the Mammalian Sensory-Motor Network
Although the biochemical heterogeneity of the mammalian sensory-motor network has been highlighted in the experiments described above, known effects of analyte suppression and enhancement suggest the use of a distinct analytical approach to validate the view established here that DRGs and adjacent nerves are chemically different. We utilized the classic GC-MS approach for this task. We investigated single DRGs and small pieces of adjacent nerves as individual samples. Small amounts of material—consisting of less than 20 (DRGs) to several hundred (distal part of sciatic nerve) milligrams of wet tissue—were collected; 206 compounds were detected in all samples (Supplementary Table S3). Possibly, because of the small amount of material available, not all compounds were observed in each sample. We characterized 58 detected compounds known to be involved in carbohydrate-energy metabolism, including sugar, ascorbate, and nucleotide metabolism, and the TCA cycle. Such compounds were detected in all regions, with glucose and lactic acid present at high levels. Nitrogen metabolism is also well represented by the presence of 31 metabolites. The classic neurotransmitters gamma-aminobutyric acid, glutamic acid, and aspartic acid were among the substances detected. Lipid metabolism is described by 96 sterols, alkenes, fatty alcohols, and fatty acids and their derivatives. Predictably, endogenous cholesterol and the stabilization agent glycerol added to the samples are present at high levels in most of the samples. Vitamin E-related compounds, including alpha-tocopherol and gamma-tocopherol, were observed in some samples, as well as five different phosphates (2-aminoethylphosphate, 3-[(2,2-dimethylpropylidene)amino]propylphosphate, monomethylphosphate, pyrophosphate (4:1), and an ester of phosphoric acid, which had the highest levels among the phosphates present).
Do the Defined Regions Have Distinct Chemical Profiles?
Intriguingly, the spinal nerve has more similarity to the dorsal root than the distal portion of the sciatic nerve collected from the rat’s leg, with the latter being chemically similar to the ventral root. The observation of the longitudinal difference in spinal-sciatic nerve structure biochemistry is in good agreement with published data demonstrating longitudinal variation in nerve morphometric parameters [63, 64, 65]. The perimeter of the axon, along with other parameters, was found to be significantly greater in the proximal part of the phrenic nerve than in the distal. However, in contrast, two nerves located in the leg, the sural nerve (a sensory branch of the sciatic nerve) and the saphenous nerve, did not demonstrate morphometric differences between their proximal and distal parts [66, 67, 68]. The close location of the DRG and spinal cord to the analyzed spinal nerve may be responsible for the biochemical differences we detected.
A high degree of chemical heterogeneity was observed in physically connected and adjacent regions of the mammalian sensory-motor system consisting of DRGs, the sciatic nerve, and ventral and dorsal roots. The chemical heterogeneity matches the functional and structural differences that the regions exhibit. The combination of several spatially-resolved MS approaches increased the information content and enabled these differences to be highlighted. We are now working to correlate the observed differences to the function of these structures, and to relate them to the specific structures and cells founds within them. A better understanding of the chemical composition of DRGs and adjacent nerves may help us to understand mechanisms of pain and develop advanced therapies for managing pain and to aid nerve regeneration.
The authors acknowledge support for this work by the National Institute of Mental Health under Award no. R21 MH100704A and the National Institute on Drug Abuse under Award no. P30 DA018310, and the National Science Foundation Division of Chemistry under grant CHE-11-11705 (with co-funding from the Division of Biological Infrastructure). The content is solely the responsibility of the authors and does not necessarily represent the official views of the funding agencies. Technical support by Xiying Wang is gratefully acknowledged.
- 7.Setou, M., Kurabe, N.: Mass microscopy: high-resolution imaging mass spectrometry. J. Electron Microsc. (Tokyo) 60, 47–56 (2011)Google Scholar
- 9.Petyuk, V.A., Qian, W.J., Chin, M.H., Wang, H., Livesay, E.A., Monroe, M.E., Adkins, J.N., Jaitly, N., Anderson, D.J., Camp II, D.G., Smith, D.J., Smith, R.D.: Spatial mapping of protein abundances in the mouse brain by voxelation integrated with high-throughput liquid chromatography-mass spectrometry. Genome Res. 17, 328–336 (2007)CrossRefGoogle Scholar
- 18.Tomlinson, L., Fuchser, J., Futterer, A., Baumert, M., Hassall, D.G., West, A., Marshall, P.S.: Using a single, high mass resolution mass spectrometry platform to investigate ion suppression effects observed during tissue imaging. Rapid Commun. Mass Spectrom. 28, 995–1003 (2014)CrossRefGoogle Scholar
- 23.National Institute of Standards and Technology. Available at: http://maldi.nist.gov. Accessed 26 Jan (2015)
- 35.Anderson, D.M.G., Mills, D., Spraggins, J., Lambert, W.S., Calkins, D.J., Schey, K.L.: High-resolution matrix- assisted laser desorption ionization-imaging mass spectrometry of lipids in rodent optic nerve tissue. Mol. Vis. 19, 581–592 (2013)Google Scholar
- 41.Devor, M.: Unexplained peculiarities of the dorsal root ganglion. Pain. Suppl 6, S27–S35 (1999)Google Scholar
- 48.Komori, N., Takemori, N., Kim, H.K., Singh, A., Hwang, S.H., Foreman, R.D., Chung, K., Chung, J.M., Matsumoto, H.: Proteomics study of neuropathic and nonneuropathic dorsal root ganglia: altered protein regulation following segmental spinal nerve ligation injury. Physiol. Genomics 29, 215–230 (2007)CrossRefGoogle Scholar
- 53.National Institute of Standards and Technology. Available at: http://www.amdis.net/index.html. Accessed 31 Jan (2015)
- 54.Wiley Registry of Mass Spectral Data. Available at: http://www.sisweb.com/software/ms/wiley.htm. Accessed 31 Jan (2015)
- 58.Wiklund, S., Johansson, E., Sjostrom, L., Mellerowicz, E.J., Edlund, U., Shockcor, J.P., Gottfries, J., Moritz, T., Trygg, J.: Visualization of GC/TOF-MS-based metabolomics data for identification of biochemically interesting compounds using OPLS class models. Anal. Chem. 80, 115–122 (2008)CrossRefGoogle Scholar
- 64.Lubinska, L., Waryszewska, J.: Fibre population of the phrenic nerve of rat: changes of myelinated fibre dimensions along the nerve and characteristics of axonal branchings. Acta Neurobiol. Exp. (Wars) 34, 525–541 (1974)Google Scholar