Article pubs.acs.org/jpr
Minocycline Effects on the Cerebrospinal Fluid Proteome of Experimental Autoimmune Encephalomyelitis Rats Marcel P. Stoop,†,# Therese Rosenling,‡,# Amos Attali,§ Roland J. W. Meesters,† Christoph Stingl,† Lennard J. Dekker,† Hans van Aken,§ Ernst Suidgeest,§ Rogier Q. Hintzen,† Tinka Tuinstra,§ Alain van Gool,∥,⊥ Theo M. Luider,† and Rainer Bischoff*,‡ †
Department of Neurology, Erasmus University Medical Center, Rotterdam, The Netherlands Department of Analytical Biochemistry, Centre for Pharmacy, University of Groningen, Groningen, The Netherlands § Abbott Healthcare Products B.V., Weesp, The Netherlands ∥ Merck Research Laboratories, MSD, Singapore ⊥ Faculty of Physics, Mathematics and Informatics, Radboud University Nijmegen, The Netherlands ‡
S Supporting Information *
ABSTRACT: To identify response biomarkers for pharmaceutical treatment of multiple sclerosis, we induced experimental autoimmune encephalomyelitis (EAE) in rats and treated symptomatic animals with minocycline. Cerebrospinal fluid (CSF) samples were collected 14 days after EAE induction at the peak of neurological symptoms, and proteomics analysis was performed using nano-LC-Orbitrap mass spectrometry. Additionally, the minocycline concentration in CSF was determined using quantitative matrixassisted laser desorption/ionization-triple-quadrupole tandem mass spectrometry (MALDI-MS/MS) in the selected reaction monitoring (SRM) mode. Fifty percent of the minocyclinetreated EAE animals did not show neurological symptoms on day 14 (“responders”), while the other half displayed neurological symptoms (“nonresponders”), indicating that minocycline delayed disease onset and attenuated disease severity in some, but not all, animals. Neither CSF nor plasma minocycline concentrations correlated with the onset of symptoms or disease severity. Analysis of the proteomics data resulted in a list of 20 differentially abundant proteins between the untreated animals and the responder group of animals. Two of these proteins, complement C3 and carboxypeptidase B2, were validated by quantitative LC-MS/MS in the SRM mode. Differences in the CSF proteome between untreated EAE animals and minocycline-treated responders were similar to the differences between minocycline-treated responders and nonresponders (70% overlap). Six proteins that remained unchanged in the minocycline-treated animals but were elevated in untreated EAE animals may be related to the mechanism of action of minocycline. KEYWORDS: multiple sclerosis, experimental autoimmune encephalomyelitis, cerebrospinal fluid, proteomics, selected reaction monitoring, minocycline
■
INTRODUCTION
pathology as well as therapeutic agents relevant for MScl can be studied. Recent clinical trials exploring the effectiveness of minocycline, a semisynthetic, second generation tetracycline analogue, showed that treatment for relapsing remitting MScl patients is promising; yet, the precise primary target of the drug remains unclear.3−8 Minocycline can effectively cross the blood−brain barrier and has been reported to demonstrate neuroprotective
Multiple sclerosis (MScl), an autoimmune disorder of the central nervous system, is the most common cause of nontraumatic disability in young adults.1 The disease is characterized by perivascular inflammatory lesions, demyelination, and axonal damage, but its etiology remains unknown.2 Experimental autoimmune encephalomyelitis (EAE), an animal model that mimics many aspects of MScl, is induced by subcutaneous injection of myelin proteins in the presence of an adjuvant. In this animal model, which can be employed in different animal species, such as mice, rats, and marmoset monkeys, MScl-related disease © 2012 American Chemical Society
Received: May 9, 2012 Published: July 6, 2012 4315
dx.doi.org/10.1021/pr300428e | J. Proteome Res. 2012, 11, 4315−4325
Journal of Proteome Research
Article
effects against a number of neurodegenerative diseases.9 While minocycline has been shown to inhibit matrix metalloproteinases,10−14 its mechanism of action in the central nervous system remains unknown, and it is unclear to which type of cells and receptors minocycline binds to exert the observed effects. One hypothesis is that the neuroprotective effect may be related to the ability of minocycline to inhibit the neurotoxic effects of matrix metalloproteinases, but other anti-inflammatory activities, such as the suppression of functional T-cell activation, may be part of the mechanism through which minocycline attenuates the inflammatory aspects of MScl, as mimicked by the EAE model used in this study. In EAE, minocycline has been shown to suppress the severity of the disease both when used as a combination treatment with atorvastatin in rats and when used as a monotherapy.15,16 Additionally, minocycline has been shown to promote remyelination in rat brain cell cultures.17 Recently, we have published our findings on the effect of EAE on the cerebrospinal fluid (CSF) proteome in an inflammatory EAE model in Lewis rats.18 In the present study, we report on the effect of minocycline treatment in this model by monitoring the CSF proteome. We combined CSF proteomics with measurement of minocycline concentrations in CSF and plasma as well as with the assessment of neurological symptoms to correlate drug concentration, the severity of the disease, and concomitant changes in the CSF proteome.
■
CSF Sampling
CSF was terminally sampled 14 days postinduction in all groups. For this procedure, the animals were euthanized with CO2/O2, and the head of the animal was held at an angle of 90° using a clamp. CSF samples were obtained by direct insertion of an insulin syringe needle (Myjector, 29G × 1/2″) via the arachnoid membrane into the Cisterna Magna. For this purpose, a skin incision was made followed by a horizontal incision in the musculus trapezius pars descendens to reveal the Arachnoid membrane. A maximal volume of 60 μL was collected per animal. Each sample was centrifuged within 20 min after sampling for 10 min at 2000g at 4 °C. After centrifugation, the supernatants were stored at −80 °C for further analysis. Previous experiments have shown that collecting up to 60 μL usually provided hemoglobinfree CSF samples as measured by Orbitrap liquid chromatography coupled tandem mass spectrometry (LC-MS/MS). Samples, which according to visual control as well as to the detection of hemoglobin-derived peptides were considered to be contaminated with blood, were excluded from the study. Immunoassay Analysis
Rat CSF samples that were collected at 14 days after induction were analyzed for levels of MMP-2 and soluble ICAM (sICAM) by enzyme-linked immunosorbent assay (ELISA). MMP-2 was detected using 96-well four-spot plates for two-plex detection of human MMP-2 and MMP-10 protein levels with claimed crossreactivity to rat orthologs (Mesoscale Discovery, cat nr K15033C-2). sICAM levels were determined using an ELISA for detection of rat sICAM (R&D systems, cat nr RIC100). CSF samples were diluted 10-fold vol/vol in corresponding assay buffers, and assays were performed following instructions from the manufacturers.
EXPERIMENTAL PROCEDURES
Induction of Acute EAE in the Lewis Rat
Male Lewis rats (Harlan Laboratories B.V., The Netherlands) were kept under normal housing conditions with water and food ad libitum. Animals weighing between 175 and 225 g at the start of the experiment were inoculated on day 0 as previously described.18 Briefly, a 100 μL saline-based emulsion containing 50 μL of complete Freund's adjuvant H37 RA (CFA, Difco Laboratories, Detroit, MI), 500 μL of Mycobacterium tuberculosis type H37RA (Difco), and 20 μg of guinea pig myelin basic protein (MBP) was injected subcutaneously in the pad of the left hind paw of animals anaesthetized under isoflurane. Next to these MBP-challenged rats, referred to as the EAE group, a control group was included, receiving the same emulsion without MBP (CFA group). Each group consisted of 30 animals and received a daily intraperitoneal administration of either minocycline [50 mg/kg in 0.01 M phosphate-buffered saline (PBS), n = 15 in both the control and the EAE group] or vehicle (2 mL/kg 0.01 M PBS, also n = 15 in both the control and the EAE group). The animals were sacrificed to collect plasma and CSF 14 days postinoculation (peak of disease in EAE group). Animals were housed in groups of three, and cages were randomized across treatments and disease duration. Disease symptoms and weights of all animals were recorded daily. Used were the following scores for motor dysfunctions: 0, healthy animal with normal curling reflex at the tail; 1, paralysis of the tip of the tail; 2, loss of muscle tone at the base of the tail; 3, low posture of hind limbs; 4, instability at hips; 5, partial hind limb paralysis; 6, complete hind limb paralysis; 7, paralysis including the midriff; 8, quadriplegia; 9, moribund; and 10, death due to EAE. The animal experiments were approved by the local Ethical Committee for Animal Experiments at Abbott Healthcare Products B. V. [Weesp, The Netherlands (study numbers: S5025.5.0002 and S5025.5.0003)].
Sample Preparation for Orbitrap-MS/MS
Protein digestion was performed in random order as follows: 10 μL of 0.1% RapiGest (Waters) dissolved in 50 mM ammonium bicarbonate was added to a sample tube containing 10 μL of CSF. The sample was reduced by adding 0.5 μL of 1,4-dithiotreitol (DTT) (0.5 M) followed by incubation at 60 °C for 30 min. After it was cooled to room temperature, the sample was alkylated with 1 μL of iodoacetamide (0.3 M) in the dark for 30 min at room temperature. To the sample, 1 μL of sequencing grade trypsin (Promega, Madison, WI, part # V5111) (1 μg/μL) (enzyme to protein ratio, 1:10−50) was added, and the sample was incubated for ∼16 h at 37 °C under slight agitation (450 rpm). Thereafter, 2 μL of trifluoroacetic acid (TFA, 50%) was added followed by incubation for 30 min at 37 °C. The samples were subsequently transferred to LC-MS sample vials. Digested rat CSF samples were measured by LC-MS/MS using an Ultimate 3000 nano LC system (Dionex, Germering, Germany) online coupled to a hybrid linear ion trap/Orbitrap mass spectrometer (LTQ Orbitrap XL; Thermo Fisher Scientific, Bremen, Germany). Five microliters of digest was loaded onto a C18 trap column (C18 PepMap, 300 μm i.d. × 5 mm, 5 μm particle size, 100 Å pore size; Dionex, The Netherlands) and desalted for 10 min using a flow rate of 20 μL/min. The trap column was switched online with the 50 cm long analytical column (PepMap C18, 75 μm i.d. × 500 mm, 3 μm particle, and 100 Å pore size; Dionex), and peptides were eluted with the following binary gradient: 0−25% eluent B for 120 min and 25− 50% eluent B for a further 60 min, where eluent A consisted of 2% acetonitrile and 0.1% formic acid in ultrapure water, and eluent B consisted of 80% acetonitrile and 0.08% formic acid in water. The column flow rate was set to 250 nL/min (oven 4316
dx.doi.org/10.1021/pr300428e | J. Proteome Res. 2012, 11, 4315−4325
Journal of Proteome Research
Article
temperature, 40 °C). For MS/MS analysis, a data-dependent acquisition method was used as follows: a high-resolution survey scan from 400 to 1800 m/z was performed in the Orbitrap {automatic gain control (AGC), 106; resolution, 30000 at 400 m/ z; lock mass set to 445.120025 m/z [protonated (Si(CH3)2O)6]19}. On the basis of this survey scan, the five most intense ions were consecutively isolated (AGC target set to 104 ions) and fragmented by collision-activated dissociation (CAD) applying 35% normalized CE in the linear ion trap. Once a precursor had been selected, it was excluded for 3 min.
quadrupole tandem mass spectrometry (MALDI-QqQ-MS/ MS) in the SRM mode using a FlashQuant Workstation containing a MALDI source [FlashLaser source (349 nm, 1000 Hz) combined with an API 4000 triple quadrupole mass analyzer (AB Sciex, Concord, Canada)] operating in the positive ionization mode. By omitting LC separation, sample analysis times were reduced to approximately 10 s per sample.20 SRM was carried out at unit resolution for the transition m/z 458.2 → m/z 441.2. Optimized instrument parameters were as follows: laser power, 55%; target plate voltage, 55 V; skimmer voltage, 0 V; CAD gas, 12 arbitrary units (collision gas, nitrogen); collision cell energy exit potential (CXP), 12 V; source gas, 15 arbitrary units; dwell time, 10 ms; laser raster speed, 1 mm/s; and collision energy (CE), 25 V. Instrument control and data acquisition were performed by Flashquant 1.0 software and Analyst 1.4.2 application software (AB Sciex, Concord, Canada). The thawed samples were diluted 1:1 (v/v) with the matrix solution [αcyano-hydroxy-cinnamic acid (6.2 mg/mL in methanol/ acetonitrile/water 36/56/8 (v/v/v), pH 2.5)], prior to spotting 0.5 μL in quadruplicate on the sample plate. Measurement of samples with known minocycline concentrations spiked was used to monitor the performance of the assay, which was performed without the use of internal standards. By spiking a known concentration of minocycline in a pooled rat plasma or CSF sample unrelated to the study and creating a dilution series from this sample, a calibration curve for minocycline in CSF was made for these SRM measurements (R2 = 0.976). The minocycline concentrations in samples from the study were subsequently calculated based on this calibration curve.
Orbitrap-MS/MS Data Processing and Analysis
The raw data were preprocessed using the Progenesis LC-MS software package (version 2.6, Nonlinear Dynamics, Newcastleupon-Tyne, United Kingdom). Peptides were identified and assigned to proteins by exporting features, for which MS/MS spectra were recorded, using the Bioworks software package (version 3.2, Thermo Fisher Scientific, Bremen, Germany; peak picking by Extract_msn, default settings). The resulting mgf file was submitted to Mascot (version 2, Matrix Science, London, United Kingdom) for identification to interrogate the UniProt database (version, 57.0; taxonomy Rattus norvegicus, 7114 sequences). Only ions with charge states between +2 and +7 were considered, and only proteins with at least two unique peptides [ion score >25 (i.e., a peptide probability cutoff value of 0.01)] assigned to them were accepted as true identifications. Modifications: carbamidomethylation of cysteine was set as fixed and oxidation of methionine as variable modification allowing a maximum of two missed cleavages. The mass tolerance for precursor ions was set to 10 ppm and for fragment ions at 0.5 Da. The cutoff for mass differences between the measured and the theoretical mass of the identified peptides was set at 2 ppm. Common contaminants such as keratins were considered in the Mascot searches, but the porcine trypsin used for digestion was not. The Mascot search results were imported back into the Progenesis software to link the identified peptides to the detected abundances of these peptides. Subsequently, the data were exported in Microsoft Excel format. In this exported matrix, the abundance of all identified peptides was listed for all samples. Only proteins identified by two or more peptides were analyzed for differential abundance between the groups (see the Supporting Information for details). It was necessary to exclude a number of animals from further analysis due to the following reasons: 1, technical failure during measurement (n = 2); 2, failure to correctly align retention times (n = 1); and 3, detection of hemoglobin (i.e., blood contamination, n = 16 from a total of 60). Each of the identified peptides was individually compared for statistically significant differences between the groups using a t test in a group-wise manner (two groups at a time). Hence, for all peptides, individual p values were calculated for all group comparisons. The abundance of proteins was considered to be significantly different in a specific comparison if at least 75% of the peptides identified for this protein had p values below 0.01 and if the difference in total abundance of the protein between the groups was larger than the difference in total protein concentration (as monitored by the UV trace) between the groups.
Quantification of Selected Proteins by LC-MS/MS in the SRM Mode
A second animal experiment was performed in which EAE was induced in the animals in exactly the same manner, only extended up to 21 days, as in the original experiment, including minocycline treatment [n = 10 animals with minocycline treatment (day 14), n = 10 animals without minocycline treatment (day 14), n = 15 animals with minocycline treatment (day 21), and n = 15 animals without minocycline treatment (day 21)]. From these animals, CSF samples were collected at days 14 and 21 and subsequently trypsin-digested. Digested CSF samples were spiked with known concentrations of stable isotope-labeled peptide standards corresponding to sequences 544−555 (EVVADSVWVDVK) and 615−621 (IWDVVEK) of complement component 3 (P01026) and sequences 155−161 (YPLYVLK) and 391−398 (YGFLLPER) of carboxypeptidase B2 (Q9EQV9) for quantification by SRM. Peptides in spiked CSF digests were separated by reversedphase chromatography on an Ultimate 3000 nano LC system (Dionex). Spiked CSF digest (10 μL) was loaded onto a C18 trap column (PepMap C18, 300 μm i.d. × 5 mm length, 5 μm particle size, and 100 Å pore size; Dionex) and washed for 5 min at a flow rate of 20 μL/min with 0.1% TFA in H2O. Next, the trap column was switched in line with the analytical column (PepMap C18, 75 μm i.d. × 150 mm length, 3 μm particle size, and 100 Å pore size; Dionex). Peptides were eluted at a flow rate of 300 nL/min with the following gradient: 0−45% solvent B in 30 min, where solvent A consisted of 2% acetonitrile and 0.1% formic acid in water and solvent B consisted of 20% water and 0.1% formic acid in acetonitrile. The separation of peptides was monitored with a UV detector (absorption at 214 nm). SRM analysis was performed on an electrospray ionization triple quadrupole (ESI-QqQ) tandem mass spectrometer (4000
Quantification of Minocycline in Plasma and CSF by MALDI-MS/MS in the Selected Reaction Monitoring (SRM) Mode
Minocycline concentrations in rat plasma and CSF samples were determined by matrix-assisted laser desorption/ionization triple 4317
dx.doi.org/10.1021/pr300428e | J. Proteome Res. 2012, 11, 4315−4325
Journal of Proteome Research
Article
Figure 1. Average CSF protein concentration of all animals at day 14, with standard deviations depicted by error bars.
group, from which hemoglobin-free CSF could be sampled, had increased neurological scores and increased CSF protein concentrations, while the other half of the animals did not, the group was split into two subgroups [animals that developed EAE (n = 5, “nonresponders” to the minocycline treatment) and animals that did not develop EAE (n = 5, “responders” to the minocycline treatment)]. Observation of the weight of the animals is a useful indicator whether MBP-injected animals are starting to develop EAE. Although minocycline itself appeared to reduce the growth rate somewhat (i.e., the gain of weight) (Figure 2), the actual loss of body weight was an indicator of EAE development. Minocycline treatment delayed the onset of EAE based on neurological scores, and disease severity was less severe in all minocycline-treated animals as compared to untreated animals. These findings were followed up in a second experiment where the animals were followed for 21 days under identical conditions. After 21 days, the animals were sacrificed, and CSF was sampled. The neurological scores observed in this 21 day experiment confirmed the previous observation that minocycline significantly delays the onset of the disease [average day on which the start of disease symptoms was observed was 11.2 ± 1.0 for untreated EAE and 13.8 ± 1.3 for EAE treated with minocycline (p = 1.6 × 10−6)]. As shown in Figure 3, minocycline significantly decreased the severity of EAE in Lewis rats [average maximum neurological score per rat for untreated EAE was 2.8 ± 0.4 and 1.7 ± 0.9 for EAE treated with minocycline (p = 0.0001)]. Additionally, minocycline treatment reduced duration [average number of days on which a neurological score greater than zero was recorded was 5.3 ± 1.0 for untreated EAE and 3.9 ± 0.8 for EAE treated with minocycline (p = 0.0002)]. Taken together, these results show that minocycline attenuates EAE in this animal model in accordance with recent clinical data in relapsing remitting MScl patients.
QTRAP; AB Sciex, Concord, Canada) in the positive ion mode. Three transitions were quantified for all peptides (EVVADSVWVDVK, transitions y6, y7, and y8; IWDVVEK, transitions y4, y5, and y6; YPLYVLK, y4, y5, and y6; and YGFLLPER, y3, y4, and y5). Data analysis was performed using Skyline (version 1.1). Concentrations of the analyte peptides were determined based on the ratio between the peak area of the analyte peptide to the peak area of the spiked isotope-labeled internal peptide standard.
■
RESULTS The focus of this study was to investigate the effects of minocycline on the CSF proteome in EAE rats in relation to its effect on neurological symptoms. All untreated EAE animals had increased CSF protein concentrations and neurological scores when compared to the control group receiving only CFA without MBP [CFA group (Figure 1)]. In addition, all untreated EAE animals had a significantly lower body weight (Table 1). Because of the fact that half of the animals of the EAE + minocycline Table 1. Characteristics of Animals and CSF Samples Taken 14 Days after Induction of EAE in Lewis Ratsa group
n
total protein concn (g/L)
weight (g)
neurological score
CFA EAE CFA + minocycline EAE + minocycline “responder” EAE + minocycline “nonresponder”
10 8 13 5
151.7 (30.0) 730.0 (204.2) 140.9 (32.0) 197.9 (34.2)
255.9 (12.3) 204.3 (21.2) 227.6 (20.0) 226.4 (10.0)
0 (0) 1.9 (1.1) 0 (0) 0 (0)
5
1008.4 (298.4)
207.6 (9.6)
1.1 (1.1)
The EAE + minocycline group is split into “responders” and “nonresponders” because 50% of the animals developed EAE, while the other 50% of the animals did not develop the disease. All values represent averages, with standard deviations in parentheses.
a
4318
dx.doi.org/10.1021/pr300428e | J. Proteome Res. 2012, 11, 4315−4325
Journal of Proteome Research
Article
Figure 2. Average weight of the animals during the 14 day experiment. Animals treated with CFA only (green, n = 10), CFA + minocycline (blue, n = 13), and the EAE animals that responded to minocycline treatment (yellow, n = 5) gained weight every day, while the animals in which EAE was induced (red, n = 8) as well as the nonresponders to minocycline treatment to EAE (purple, n = 5) showed a decrease in weight eventually.
Figure 3. Bar chart of the average neurological scores of animals sacrificed at day 21. Animals with EAE (orange, n = 15) had higher neurological scores than animals with EAE that were treated with minocycline (green, n = 15). In addition, disease development was delayed in the minocycline-treated animals. Error bars depict the standard deviation for the average of each column.
effect of minocycline on levels of sICAM protein (EAE, 18.53 ± 3.08 ng/mL; EAE + minocycline, 28.03 ± 5.68 ng/mL; p value
