Proteomics of Neuropathic Pain: Proteins and Signaling Pathways

for repeat recycling. Developing a practical polymer that can break down to its monomer and then be repolymerized over and... BUSINESS CONCENTRATE...
1 downloads 0 Views 364KB Size
Download PDF
Article pubs.acs.org/jpr

Proteomics of Neuropathic Pain: Proteins and Signaling Pathways Affected in a Rat Model Ping Sui,*,† Hiroyuki Watanabe,‡ Michael H. Ossipov,§ Georgy Bakalkin,‡ Konstantin Artemenko,†,§ and Jonas Bergquist*,†,§ †

Analytical Chemistry, Department of Chemistry − BMC and SciLifeLab and ‡Molecular Neuropsychopharmacology, Department of Pharmaceutical Biosciences, Uppsala University, P.O. Box 599, Husargatan 3, 75124 Uppsala, Sweden § Department of Pharmacology, University of Arizona Health Sciences Center, 1501 North Campbell Avenue, Tucson, Arizona 85724, United States S Supporting Information *

ABSTRACT: The myriad proteins may be involved in the mechanisms underlying the development and maintenance of neuropathic pain, an extremely disabling condition that originates from pathology of the nervous system. To address the mechanisms, we here analyzed proteins and cellular networks in the dorsal spinal cord mediating pain processing in a well-established rat model of neuropathic pain induced by spinal nerve ligation (SNL). Labelingbased proteomic methods together with high-resolution mass spectrometry for proteome analysis were applied. 38 proteins including synapsin 1 and microtubule-associated protein 2 were identified as differently expressed in the SNL group. Pathway analysis suggests that maladaptive changes in the levels of these proteins may contribute to abnormal synaptic transmission and neuronal intracellular signaling underlying the onset and development of neuropathic pain. KEYWORDS: mass spectrometry, protein quantification, neuropathic pain, dimethyl labeling, rat spinal cord

1. INTRODUCTION Neuropathic pain (NP) is defined as a pain caused by a primary lesion or dysfunction of the nervous system.1 It could be caused by numerous diseases, such as diabetes, herpes zoster infections, trauma, cancer, and neurotoxins,2 causing damage to central or peripheral nervous tissue. Currently, several candidate compounds have been proposed for treatment. However, the mechanism of how these drugs relieve pain is not fully understood. Studies using animal models of NP provided several hypotheses about proteins involved in neuropathy.3 An experimental model of mononeuropathy, spinal nerve ligation (SNL), was reported by Kim and Chung.4 In this model, L5 and L6 nerves of the rat were tightly ligated. An exquisite sensitivity of the hindpaw is introduced within 1 day after surgery and gradually increased in the following several days and lasted for months. It developed symptoms of NP as hyperalgesia and allodynia,5 which are the characteristics of stimulus-evoked pain in animal models.6 Studying the global changes in protein expression level could lead to the identification of pain associated proteins and how their composition changes in the central nervous system during the generation or maintenance of pain. It is very important for understanding the molecular mechanisms of NP and selecting new drug targets or diagnostic markers. © 2014 American Chemical Society

Western blot and immunohistochemistry are commonly used for protein quantification in normal and pathological tissues.7,8 The limitations of these techniques are their inherent ability to process one or several proteins at the time.9 Development of proteomics techniques solved this problem by allowing simultaneous analysis of hundreds or thousands of proteins. Besides protein quantification, proteomics also provides information about protein structure, posttranslational modification, and protein−protein interaction.10 Together with systems biology, proteomics can be used to explore protein interaction network for studying pathophysiology.11 In NP research, most of the studies in proteomic were done using 2-D polyacrylamide gel electrophoriesis technique (2DPAGE) for protein quantification.2 With 2D-PAGE, proteins extracted from samples were first separated in a sodium dodecyl sulfate polyacrylamide gel according to their properties, that is, isoelectric point, molecular weight, and solubility. Then proteins in the gel were visualized by chemical stains or fluorescent markers, and the intensity of the protein stain is used for quantification. Protein identification was performed by tandem mass spectrometer (MS/MS) coupled to a liquid Received: March 10, 2014 Published: June 16, 2014 3957

dx.doi.org/10.1021/pr500241q | J. Proteome Res. 2014, 13, 3957−3965

Journal of Proteome Research

Article

L6 spinal nerves were tightly ligated distal to the dorsal root ganglion with a 4−0 silk suture on the right side, and the incision was closed. Sham control rats underwent the same surgery and handling as the experimental animals but without SNL. No surgeries were done for Naive rats. Rats were sacrificed by decapitation on the 10th day after surgery, immediately after behavioral testing. The spinal cord was dissected by cutting the vertebral column at the base of the skull and just above the hip bone. A syringe was inserted at the lower opening, and the spinal cord was flushed out with icecold saline onto a glass plate placed on ice. For all rats, the lumbar enlargement region was immediately isolated and divided into left and right dorsal and ventral segments, and the right dorsal segments were taken for analyses.

chromatography (LC) system. The protein separation of 2DPAGE has a good sensitivity, linearity, and dynamic range. It is possible to separate thousands of proteins, but it is limited to abundant and soluble proteins.12 In fact, many proteins in the neuronal system are hydrophobic,13 that is, membrane receptor proteins. Mass-spectrometry-based proteomics allows gel-free approaches to protein quantification, which, unlike 2D-PAGE, could rapidly identify and quantify the heavy, low abundant, and both hydrophobic and hydrophilic proteins. Several early proteomic studies identified proteins that associate with the NP in animal models.2,9,14 Changes in the levels of these proteins depend on specific pain model and time course of pain development. So far, a general pattern of proteomic changes and their role have not been yet established. However, rearrangements of multiple protein pathways catalyzed by the ubiquitine−proteasome system have been demonstrated to be a critical molecular event underlying development and maintenance of neuropathic and inflammatory pain.15−17 Recently, we developed an efficient method for the proteome quantification in spinal cord samples that is robust and reliable from protein extraction and labeling to protein identification and quantification.18 In this method, isotopic dimethyl labels are introduced into proteome digests, followed by LC−MS/MS analysis.19 Common reagents are used in this approach that make it an economic option for large-scale quantitative analyses.20 In this study, we examined the protein component in the dorsal spinal cord of SNL rats compared with that of two control groups including Sham-operated and Naive animals. Those two control groups were commonly introduced in NP studies;21 however, it is the first time to use two controls in the study of global protein expression in NP with quantitative proteomics methods. We identified several proteins showing significant changes in their levels in the SNL group compared with other two groups. Furthermore, protein pathway database analysis predicted protein networks and proteins affected by NP, which cannot be identified by MS analysis alone.

2.3. von Frey and Hot-Plate Test

Responses to noxious thermal stimuli were determined by the hot-plate test. Latency to withdrawal and licking of a hindpaw was determined by placing the animal on a metal plate maintained at 52 °C. A maximal cutoff of 40 s prevented tissue damage. Importantly, all rats with nerve injury always responded by lifting or licking the hindpaw ipsilateral to the nerve injury, whereas animals with sham surgery did not demonstrate any preference in the hindpaw response. The hotplate latencies of the nerve-injured rats were consistently significantly lower than those of the sham-operated animals. Paw withdrawal thresholds to light tactile stimuli were measured by probing the hindpaw with a series of calibrated von Frey filaments (0.41, 0.70, 1.20, 2.00, 3.63, 5.50, 8.5, and 15 g) applied to the plantar aspect of the hindpaw according to the “up-down” method as described previously.23 Sham-operated rats did not respond to the cutoff filament (15 g). This strength of tactile stimulus is not considered to be nociceptive.23 Consequently, it was not possible to quantify a reduced sensitivity, or “hypoesthesia”, to non-noxious tactile stimuli. Higher strengths of von Frey filaments had the effect of lifting the animal’s hindpaw from the grid, thus necessitating the 15 g cutoff.24,25 Evidence of tactile allodynia was confirmed in rats used for biochemical assessments. Mechanical allodynia and thermal hyperalgesia were determined by a significant (p < 0.05) reduction in response values from pretreatment baseline values. In addition, the animals were observed for primary overt effects on behavior, such as exploratory behavior, gait, and posture.

2. EXPERIMENTAL PROCEDURES All Chemicals, if not specifically stated, were purchased from Sigma-Aldrich, Steinheim, Germany. 2.1. Sample Preparation

All experimental procedures were performed in accordance with the policies and recommendations of the International Association for the Study of Pain and the National Institutes of health guidelines for the handling and use of the laboratory animals and were approved by the Institutional Animal Care and Use Committee of the University of Arizona. Thirty male Sprague−Dawley rats (Harlan, Indianapolis, IN) weighing between 225 and 300 g were maintained in cages in a climate-controlled room on a 12 h light/dark cycle with ad libitum access to food and water.

2.4. Protein Extraction

2 to 3 mg of homogenized rat dorsal spinal cord tissue powder of ipsilateral dorsal segment of lumbar spinal cord was dissolved in 100 μL of lysis buffer containing 8 M urea and 1% octyl-β-Dglucopyranoside in 10 mM Tris (Merck, Darmstadt, Germany), pH 8.0. The tissues were sonicated for 60 min at 0 °C in water bath (Elma, Germany) and then centrifuged for 10 min at 15 000g at 4 °C (Sigma, Germany). Extracted proteins in the supernatant were gently collected. The concentration of extracted proteins was determined using Bradford protein assay (Bio-Rad laboratories, Hercules, CA) according to the manufacturer’s protocol.

2.2. SNL-Induced Neuropathic Pain Model

Three groups of rats, Sham-operated, SNL, and Naive rats were used in this study. There were 10 rats in each group. Peripheral nerve injury was produced by tight ligation of the L5 and L6 spinal nerves, as described by Kim and Chung,4 and as performed routinely in our laboratories.17,22 Anesthesia was induced with 2% isoflurane in air at 2 L/min and maintained with 0.5% isoflurane in air. The dorsal vertebral column from L4 to S2 was exposed, and the L5 and L6 spinal nerves of the right hindpaw were identified and carefully isolated. The L5 and

2.5. On-Filter Digestion

A 20 μg protein aliquot of rat spinal cord extraction reduced (15 mM dithiothreitol in 56 °C for 30 min) alkylated (33 mM iodoacetamide in the dark for 30 min). samples were transferred onto 3 kDa cutoff filters (Pall 3958

was and The Life

dx.doi.org/10.1021/pr500241q | J. Proteome Res. 2014, 13, 3957−3965

Journal of Proteome Research

Article

Table 1. Sample Labeling Setup for Relative Quantification Detected With LC Coupled to FT−ICR MSa dimethyl labels mixture samples

a

label

light

medium

formaldehyde isotope cyanoborohydride isotope mixture 1 mixture 2 ... mixture 10

CH2O NaBH3CN Sham-01 Sham-02 ... Sham-10

CD2O NaBH3CN SNL-01 SNL-02 ... SNL-10

heavy 13

CD2O NaBD3CN Naive-Pooled Naive-Pooled ... Naive-Pooled

Each mixture samples was analyzed in LC−MS/MS system twice.

Figure 1. Development of mechanical allodynia and thermal hyperalgesia in the Sham and SNL groups. The threshold and response was tested presurgery and on the 6th and 10th days postsurgery for both groups. In von-Frey test, the cutoff with 15 g of filament was introduced; only one rat in Sham group withdrew its paw at 10.4 g of filament, while the rest of the rats in both groups showed paw withdrew above the 15 g threshold before surgery. Data obtained prior to surgery are not presented as a box plot because of the introduced cutoff threshold. The average value is shown as the diamond mark, and the median value is the straight line across the box. t test was used to evaluate significance of differences between Sham and SNL groups. *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001.

Sciences, Ann Arbor, MI) and centrifuged at 14 000g for 15 min, then additional buffer exchanging was performed in three steps as follows: (i) 50 μL of 2% acetonitrile in 100 mM triethylammonium bicarbonate (TEAB); (ii) 50 μL of 50% acetonitrile in 100 mM TEAB; and (iii) 50 μL of 100 mM TEAB. The proteins were digested on-filter with 1 μg of trypsin (Roche Diagnostics, Mannheim, Germany) in 50 μL of 100 mM TEAB at 37 °C for 90 min. The digests were spun through the filter at 14 000g. 50 μL of 0.1% formic acid was added onto the filter. Samples were spun at 14 000g for 30 min to quench the digestion. The two last eluates were combined and dried in SpeedVac (Thermo Scientific, Waltham, MA).

solution in water was added to each light and medium labeled sample, and the same amount of 0.6 M NaBD3CN (ISOTEC) was added to each pooled Naive sample for heavy labeling, and after being shaked for 1 h at room temperature, the reaction was terminated by adding 16 μL of 1% ammonia solution (Merck) and then 8 μL of 5% formic acid (Merck). The samples labeled with light, medium, and heavy isotopes were mixed together and cleaned up using C18 SPE columns (1 mL, 50 mg capacity, Biotage, Uppsala, Sweden) according to manufacturer’s instruction. After desalting, samples were dried in SpeedVac and redissolved in 0.1% trifluoroacetic acid prior to nano-LC−MS/MS.

2.6. Dimethyl Labeling

2.7. Nano-LC−MS/MS

Five μL aliquot containing ∼2.5 μg of each labeled sample mixture was injected into a nanoLC−MS/MS system consisting of (i) 1100 series HPLC (Agilent Technologies, Waldbronn, Germany) and (ii) LTQ-FT Ultra mass spectrometer (Thermo Fisher Scientific, Bremen, Germany). The separation was carried out on a 150 mm × 75 μm analytical column packed inhouse with Reprosil-Pur C18-AQ, 3 μm resin (Dr Maisch, Ammerbuch-Entringen, Germany). The peptides were first loaded onto column with 98% buffer A (0.5% acetic acid in water) and 2% buffer B (acetonitrile with 0.5% acetic acid, and 10% water) at 500 nL/min. Gradient elution was from 4% B to 50% B in 40 min at a flow rate of 200 nL/min. Mass spectrometric analysis was performed using unattended datadependent acquisition in which the mass spectrometer

26

Dimethyl labeling was performed as described elsewhere. In brief, 10 μg of tryptic peptides of rat spinal cord from individual rat was reconstituted in 50 μL of 100 mM TEAB. The aliquots of tryptic peptides obtained from Naive rat samples were pooled together and used as internal control for further normalization in the downstream analysis. (See Table 1.) All samples should have the same protein concentration at this point. Samples from Sham-operated, SNL and Naive rats (10 μg proteins/50 μL) were mixed with 4 μL of 4% regular formaldehyde−CH2O or deuterated formaldehyde−CD2O (ISOTEC, Miamisburg, OH) or deuterated and 13C-labeled formaldehyde−13CD2O (Sigma-Aldrich, St. Louis, MO) and marked as light, medium, and heavy, respectively (Table 1). After brief vortexing, 4 μL of 0.6 M NaBH3CN (ISOTEC) 3959

dx.doi.org/10.1021/pr500241q | J. Proteome Res. 2014, 13, 3957−3965

Journal of Proteome Research

Article

Figure 2. Variation distribution in Sham and SNL rat groups. (a) PCA plots of behavior data set. (b) PCA plots of relative protein levels.

3.1. Mechanical Allodynia and Thermal Hyperalgesia Induced by SNL

automatically switched between acquiring a high-resolution survey mass spectrum in the FTMS (resolving power 50 000 fwhm at m/z 524) and five consecutive low-resolution, collision-induced dissociation fragmentation of up to five of the most abundant ions in the ion trap. Automatic gain control (AGC) was set to 106 (FTMS, full MS) and 104 (LTQ, MSn). CID was performed with helium as a collision gas (pressure 0.1 Pa), and normalized collision energy was set to 35.0.

In this work, we used a well-established rat NP model that was applied in many studies.1,2 Figure 1 shows that behavioral effects of SNL and Sham operation correspond to the data previously published by several groups.27,28 Von-Frey and hotplate tests were performed on Sham and SNL rats before the surgery and also on the 6th and on the 10th days after the surgery. Before surgery, rats in the Sham and SNL groups exhibited similar baseline thresholds to the mechanical and thermal stimuli. The rats with SNL showed significant reductions in mechanical threshold and thermal responses on both the 6th and 10th days, indicating behavioral signs of the NP. The sham group of rats showed slight nonsignificant increases in mechanical and thermal responses.

2.8. Data Analysis

The raw data files were exported to MGF format using an inhouse written program. Peptides and proteins were identified by Mascot V2.2.2 (Matrix Science, London, U.K.) against SwissProt database (version 51.6) with a precursor mass tolerance set to 10 ppm, a fragment ion mass tolerance of ±0.7 Da, and strict trypsin specificity allowing up to two missed cleavages. Taxonomy was set to Rattus. Fixed modification was carbamidomethyl, and variables modifications were oxidation, deamidated, dimethyl (K)/(N-term), dimethyl:2H(4) (K)/(Nterm),and dmethyl:2H(6)13C(2) (K)/(N-term). Proteins were only considered to be positively matched if they passed the more stringent MudPIT MASCOT ion scoring (p ≤ 0.05) and at least one peptide passing the required bold red criteria. Relative quantification results of identified proteins were calculated through MSQuant v2.0. Principal component analysis (PCA) was processed with Qlucore Omics Explorer 2.3. Protein pathway analysis was performed using IPA pathway searching engine from Ingenuity System, Redwood City, CA.

3.2. Principal Component Analysis Analysis

Differences between the SNL and Sham groups in nociceptive behavior and protein profiles established by LC−MS were examined by the PCA. PCA was performed with two data sets that are (i) two behavior tests and (ii) LC−MS protein quantification results. First, measurements by von-Frey and hotplate tests on the 6th and 10th days postsurgery for each rat in the Sham and SNL groups were submitted into PCA as variables. In the PCA score plot, each rat was presented as one point and the rats with similar behavior responses will be located closer. Second, protein expression levels in the Sham and SNL groups were also processed in PCA after their experimental normalization using samples from Naive animals; that is, in mixture 1, ratios of Sham/Naive and SNL/Naive of each protein were calculated and used as the quantification results for rat no. 1 in Sham group and rat no.1 in SNL group, respectively. Because each labeled mixture was injected twice (see Table 1), we first calculated the average ratio for each protein. If protein was found only in one replicate, this single ratio was kept. Altogether 498 proteins were identified with 153 of them present in at least 9 rats in each group. Expression ratios of these 153 proteins for each rat were submitted to PCA. For the proteins quantified in 9 out of 10 labeled mixtures, the missing values were assigned as predicted by software. In the PCA score plot, the position of each rat depended on the protein expression ratios. The rats with similar proteome formed a

3. RESULTS To identify proteins involved in NP, we compared protein profiles in the dorsal lumbar spinal cord between the SNL and Sham operation groups. The Sham rats were operated similarly to the SNL animals, but spinal nerve was not ligated; therefore, this group was considered to be a proper control for neuropathic animals. However, sham operation may induce inflammation and stress that, in turn, affects the protein levels. Therefore, in addition, the Naive rats were included in the analysis to control the effects of sham operation. Two-way comparison may allow identification of proteins that are specifically associated with chronic NP but not with acute pain. 3960

dx.doi.org/10.1021/pr500241q | J. Proteome Res. 2014, 13, 3957−3965

Journal of Proteome Research

Article

(Table 2). The most stringent correction for multiple comparisons, the Bonferroni testing with a new p value (p < 0.05/153 = 0.0003) identified three proteins, SYN1 (Synapsin 1), MAP2 (Microtubule-associated protein 2), and LDHA (Lactate dehydrogenase A), which remain significantly different (Table 2). SYN1 and MAP2 are substantially down-regulated in SNL rats with no gross changes in the Sham group. Change in LDHA is highly significant for the SNL group compared with Sham group. However, this protein is also affected by surgery, inflammation, or acute pain. Consistently, LDH enzymes were reported to be associated with inflammation.29 The biochemical validations of protein expression were not performed on SYN1 and MAP2. They have been already studied by the other groups and our results confirmed their observation.30,31

cluster in the plot. As the result shown in Figure 2, the rats in Sham and SNL groups demonstrated clear separation. The SNL group demonstrated lower spread compared with the Sham group on both PCA plots. This may be due to strong effects of the SNL inducing NP under which interindividual differences in nociceptive and non-nociceptive responses as well as in protein expression profiles could become less pronounced. 3.3. Proteins Affected by SNL

Differences in expression level of each of 153 proteins between the two animal groups were examined by t test with two-tailed distribution using their logarithm ratios. Data sets with Sham/ Naive and SNL/Naive ratios used in the PCA were processed. Difference was considered to be significant if the p value was