Article pubs.acs.org/jpr
A Comparative Proteomic Study of Human Skin Suction Blister Fluid from Healthy Individuals Using Immunodepletion and iTRAQ Labeling André C. Müller,† Florian P. Breitwieser,† Heinz Fischer,‡ Christopher Schuster,§ Oliver Brandt,§ Jacques Colinge,† Giulio Superti-Furga,† Georg Stingl,§ Adelheid Elbe-Bürger,*,§ and Keiryn L. Bennett*,† †
CeMM Research Center for Molecular Medicine of the Austrian Academy of Sciences, Vienna, Austria Department of Dermatology, Research Division of Biology and Pathobiology of the Skin, Medical University of Vienna, Vienna, Austria § Department of Dermatology, Division of Immunology, Allergy and Infectious Diseases, Medical University of Vienna, Vienna, Austria ‡
S Supporting Information *
ABSTRACT: Aberrations in skin morphology and functionality can cause acute and chronic skin-related diseases that are the focus of dermatological research. Mechanically induced skin suction blister fluid may serve as a potential, alternative human body fluid for quantitative mass spectrometry (MS)-based proteomics in order to assist in the understanding of the mechanisms and causes underlying skin-related diseases. The combination of abundant-protein removal with iTRAQ technology and multidimensional fractionation techniques improved the number of identified protein groups. A relative comparison of a cohort of 8 healthy volunteers was thus recruited in order to assess the net variability encountered in a healthy scenario. The technology enabled the identification, to date, of the highest number of reported protein groups (739) with concomitant relative quantitative data for over 90% of all proteins with high reproducibility and accuracy. The use of iTRAQ 8-plex resulted in a 66% decrease in protein identifications but, despite this, provided valuable insight into interindividual differences of the healthy control samples. The geometric mean ratio was close to 1 with 95% of all ratios ranging between 0.45 and 2.05 and a calculated mean coefficient of variation of 15.8%, indicating a lower biological variance than that reported for plasma or urine. By applying a multistep sample processing, the obtained sensitivity and accuracy of quantitative MS analysis demonstrates the prospective value of the approach in future research into skin diseases. KEYWORDS: skin suction blister fluid, iTRAQ, Orbitrap, gel-free, depletion, body fluid proteomics
■
INTRODUCTION The skin accounts for approximately 15% of the total adult body weight, is the largest organ and exerts multiple crucial functions. The skin forms a protective barrier, shields muscles, internal organs, and body fluids from bacteria, viruses, ultraviolet light and other environmental aggressors. Additionally, skin plays a crucial role in the defense mechanism of the body that is controlled by several types of immune cells. It associates tissues of various origins that are organized in three layers including the outermost epidermis (ectodermal), dermis and the hypodermis (mesenchymal). The epidermis and dermis are connected by the dermal−epidermal junction forming a complex basement membrane including the lamina lucida layer.1 A major goal of basic skin research is to provide early detection and optimal treatment of widespread skin diseases © 2012 American Chemical Society
through an improved understanding of the underlying physiology and pathology. Mass spectrometry (MS)-based technologies have been widely used for large-scale characterization of complex human body fluids, e.g., plasma or serum, with the focus on diseaserelated marker proteins.2 “Proximal” fluids, derived from the extracellular milieu of tissues such as cerebrospinal fluid, bronchoalveolar lavage fluid or skin suction blister fluid (SBF), represent potential alternative sources for clinically relevant proteomic research.3 The method to induce suction blisters on the skin was established in 1968 by Kiistala4 and is based on the finding that the application of low pressure on the skin surface for 2−3 h results in the separation of the epidermis from the Received: March 1, 2012 Published: May 14, 2012 3715
dx.doi.org/10.1021/pr3002035 | J. Proteome Res. 2012, 11, 3715−3727
Journal of Proteome Research
Article
Figure 1. Schematic overview. (A) Skin blisters were induced by prolonged application of suction using vacuum chambers with 5 mm hole plates mounted onto the forearm. Suction blister fluid (SBF) of visually blood uncontaminated blisters were aspirated and pooled into an eppendorf tube containing protease inhibitors. (B) For the comparison of two depletion strategies, SBF from one control volunteer was processed, and tryptic peptides were chemically labeled with iTRAQ reagent. Analysis was performed using two-dimensional reversed phase chromatography (2D-RPRP) at different pH in combination with a hybrid LTQ-Orbitrap XL mass spectrometer employing higher energy collision-induced dissociation (HCD) only. (C) For the assessment of biological and technical variability of the technology, 8 healthy controls were subjected to the same procedure, and equal amounts of tryptic peptides were either used for an individual comparison with iTRAQ 8-plex or pooled using iTRAQ 4-plex reagent. Analysis was performed as before, but with a modified hybrid CID-HCD methodology using normal collision-induced dissociation (CID) in the linear ion trap for identification and HCD at elevated energies for iTRAQ quantitation for improved sensitivity and accuracy.
dermis along the lamina lucida. The fluid from the freshly formed blister compartments corresponds to the interstitial tissue fluid of the dermis5 and has been routinely used in studies on drug pharmacokinetics and for the assessment of inflammatory mediators during skin inflammation.6 The composition of SBF is not well-characterized; however, it has been shown that some serum proteins and lipid constituents are up to 5-fold lower in this fluid compared to serum. This is similar to results obtained for peripheral lymph fluid.7 In 2007, Kool and colleagues8 were the first investigators to conduct a large-scale, MS-based proteomic analysis of SBF in the search for novel biomarkers. When directly compared to serum, the data showed a comparable protein pattern by two-dimensional gel electrophoresis (2D-PAGE). The subsequent proteomic analysis confirmed a similar relative abundance of the 10−20 most abundant proteins. Additionally, specific classes of proteins (e.g., cell leakage) were predominantly, or exclusively, identified in SBF. The study systematically demonstrated that MS-based proteomics is a rapid and powerful tool to identify a considerable proportion of the skin blister fluid proteome. Despite stating the prospective value in terms of proteomic targeting and monitoring of proteins implicated in skin-related diseases, no dermatological comparative study of a defined pathological condition was undertaken. To date, only a handful of comparative clinical studies employing different degrees of MS-based proteomics have been reported. Patients with skinrelated diseases such as plaque psoriasis,9 mycosis fungoides,10 and atopic dermatitis11 were employed in the studies. A major obstacle in the proteomic analysis of human body fluids is the very high complexity and the dynamic range of
protein concentrations that can span 10 orders of magnitude; e.g., 95% of the total serum proteins is comprised of only 10 proteins.12 To allow the detection of low-abundance proteins, multidimensional fractionation methods are necessary to minimize the interference of abundant proteins.13 It has been demonstrated that the sieve function of the vascular wall of the microcirculatory network beneath the epidermal basement membrane remains intact during blister formation, and the degree of protein diffusion from plasma is reciprocal to the molecular weight.14 As a consequence, the protein content of SBF was found to be around 30% of that of plasma and therefore similar to results reported for peripheral lymph fluid. Common fractionation techniques include antibody-based depletion methods and resin-based multidimensional liquidchromatographic approaches in order to improve the dynamic range of detection by mass spectrometry as shown for SBF.8 Commercially available antibody-based depletion strategies vary in the number of targeted proteins and the origin, specificity and cross-reactivity of used antibodies. Hence, depletion efficiencies will vary, and thus prior to undertaking clinical studies, it is advisible to evaluate the planned approach and applicability. Here, the two commercial depletion spin columns MARS human 6 (top 6) and the Seppro IgY14 (top 14) were compared. Both remove albumin, IgG, IgA, α1-antitrypsin, transferrin and haptoglobin. In addition, the latter removes IgM, α2-macroglobulin, fibrinogen, complement C3, α1-acid glycoprotein, apolipoprotein A-I/II (HDL) and apolipoprotein B (LDL). The major differences therefore are the depth of protein removal and the specificity of the employed polyclonal antibodies. 3716
dx.doi.org/10.1021/pr3002035 | J. Proteome Res. 2012, 11, 3715−3727
Journal of Proteome Research
Article
WI); protease inhibitor cocktail (Roche Diagnostics, Mannheim, Germany).
Different proteomic approaches exist to measure quantitative changes in the investigated samples that vary in the sensitivity of the analysis and accuracy of quantitation.15 So far, comparative studies of SBF proteomes were performed by 2D-PAGE9,10 or semiquantitatively by using unique peptide counts8 or spectral counts.11 It is well-known that 2D-PAGE has certain limitations with highly complex samples, and reproducibility can achieve a coefficient of variation (CV) of 20%.16 Semiquantitative label-free LC−MS approaches have provided a high level of coverage of the proteome but show lower accuracy with respect to quantitation.15 The use of isobaric tags for relative and absolute quantitation (iTRAQ)17 in MS has gained popularity and offers the advantages of multiplexing capabilities (up to eight channels), enhancing low intensity ions,18 and compatibility with clinical samples. Stable mass tags have been successfully utilized in plasma19 and cerebrospinal fluid20 proteomics, but to date, no report has been published where this methodology has been applied to skin SBF. As with any technology, there are certain shortcomings, especially when combined with ion trap-based mass analysers such as the LTQ Orbitrap XL.21 In brief, the requirement for the higher energy collision-induced dissociation (HCD) cell to efficiently trap and detect the iTRAQ reporter ions comes at the cost of reduced sensitivity, retardation of the duty cycle and an increase in total scan time. Also, the larger iTRAQ 8-plex tag shows unfavorable kinetic consequences for the MS2 fragmentation of peptides and together with interfering tag-derived peaks causes a reduced identification rate.21 As with many quantitative assays, especially when combined with such a complex technology as MS-based proteomics, there is a certain impact of technical noise and variability. For that reason, sound statistical models are necessary for a robust and sensitive determination of quantitative differences especially with respect to interindividual differences. To date no gel-free quantitative MS-based methodology for a comparative study of skin SBF proteomes has been described that provides high sensitivity in the range of low-abundance proteins with concomitant monitoring of relative concentrations with high accuracy. Therefore, the aims of this work were to first assess two commercially available depletion spin cartridges in terms of improved dynamic range of mass spectrometric analysis and incorporation into a quantitative proteomic workflow using iTRAQ labeling (Figure 1). Second, the preferred strategy was then applied to a cohort of eight healthy individuals to evaluate technical and biological variability in a multiplexed fashion and investigate possible gender differences. Third, an enzymatic activity assay with undepleted SBF was performed in order to validate quantitative data obtained for a subgroup of ribonucleases (RNases), recently recognized as an important part of the innate immune response to invading pathogens via skin infection.22 The potential value of the presented quantitative mass spectrometric-based platform for clinical research of skin-related diseases will be discussed.
■
Skin Suction Blister Formation, Sample Collection, and Preparation
Skin suction blister fluid was collected from 8 healthy volunteers by the method developed by Kiistala in 1968.4 The study was approved by the local ethics committee and conducted in accordance with the declaration of Helsinki Principles. Written informed consent was obtained from all volunteers. Inclusion criteria were as follows: adult, Caucasian, healthy volunteers (25−45 years) of both sexes. Exclusion criteria were as follows: oral drug intake (e.g., antibiotics), topical medication or exposure to UVB light 7−10 days prior to the start of the study; pregnancy; immunosuppression; severe systemic diseases or other diseases involving the skin. The cohort consisted of 4 females with a mean age of 34 years (range 28−42 years) and 4 males with a mean age of 32 years (range 26−41 years). Skin suction blisters were formed using a low pressure instrument from Electronic Diversities (Ridge Road, Finksburg, MD) by mounting 2 sterile, 5-hole (5 mm diameter per hole) skin suction plates onto the forearm of the volunteers. Low pressure (150−200 mmHg) was applied over a time period of 2−3 h, and strength was adjusted according to blister size in order to prevent blood vessel trauma through excessive mechanical stress. Fluid from intact and hemoglobin-free blisters were collected using a Micro-Fine syringe (BD Biosciences, Maryland, US) and pooled into an eppendorf tube containing 7 μL of 25× protease inhibitor cocktail (Roche Diagnostics, Mannheim, Germany), mixed thoroughly and stored on ice. In addition, blood samples were taken from each volunteer and collected in plasma and serum tubes (Vacuette, Greiner Bio-One GmbH, Austria). Immediately after collection, SBF samples (115−180 μL) were centrifuged at 1600g for 15 min at 4 °C in a refrigerated centrifuge (Eppendorf, Hamburg, Germany) to pellet any intact cells. The supernatant was removed and centrifuged at 16 000g for 15 min at 4 °C to remove any cellular organelles and debris from apoptotic cells. The supernatant was removed, and aliquots were snap-frozen in liquid nitrogen and stored at −80 °C until required. Plasma tubes were centrifuged at 1700g for 15 min, the supernatant removed, and 1 mL aliquots mixed with 40 μL of 25× protease inhibitor cocktail. Aliquots were snap-frozen in liquid nitrogen and stored at −80 °C. Protein Quantitation and Depletion
Protein concentration was determined using the PIERCE protein assay (ThermoFisher Scientific, Waltham, MA) according to the instructions provided by the manufacturer. For the comparison of the two depletion strategies, either 600 or 800 μg total protein from the SBF of a single volunteer was used for the depletion. This was achieved using the MARS human 6 (Agilent Technologies, Santa Clara, CA) or the Seppro IgY 14 (SIGMA-Aldrich, St. Louis, US) spin column kit, respectively. Depletion was performed according to the instructions supplied by the manufacturer. For the comparison of the 8 individuals, 700 μg of total protein was used for a sequential depletion with the Seppro IgY 14 kit. In all cases, the protein concentration in the flow through containing the depleted sample was estimated using the Qubit fluorometer and Quant-iT protein assay allowing an approximate calculation of the yield of depletion.
MATERIALS AND METHODS
Materials
Materials used in this study included iodoacetamide, dithiothreitol, triethylammonium bicarbonate (TEAB), formic acid, Tris-HCl, bromophenol blue, toluidine Blue O (SIGMAAldrich, St. Louis, MO); trypsin (Promega Corp., Madison, 3717
dx.doi.org/10.1021/pr3002035 | J. Proteome Res. 2012, 11, 3715−3727
Journal of Proteome Research
Article
In-Solution Tryptic Digestion
(ii) an improved hybrid collision-induced dissociation (CID)HCD method was employed as described,25 whereby a CID fragmentation in the linear ion trap (LTQ) was followed by a HCD fragmentation of the same precursor ion at an elevated energy of 80% CE in the HCD cell of the LTQ Orbitrap XL.
For the comparison of the depletion strategies, 60 μg of depleted protein was used. For the comparison of 8 healthy volunteers, all the protein was used for sample preparation. Samples were denatured in 6 M urea, reduced with dithiothreitol, alkylated with iodoacetamide and digested overnight at 37 °C with 2.5 μg of modified porcine trypsin (Promega Corp., Madison, WI) after dilution of urea below 2 M. The latter samples were lyophilized before digestion in order to reduce the volume. Tryptic peptides were acidified and concentrated with C18 solid phase extraction SPE spin columns (#SEM SS18V, The Nest group, Southborough, MA), and eluates were dried in a vacuum concentrator. Samples were resuspended in 500 mM TEAB buffer, pH 8.2, prior to labeling with iTRAQ reagent (ABI, Framingham, MA).
Data Analysis
The acquired data were processed with Bioworks V3.3.1 SP1 (ThermoFisher, Manchester, UK). The .dta files were extracted from the .raw files with the extract_msn.exe program with the following settings: −F1 −L0 −B600 −T4500 −M0.01 −C2 −S1 −I5 −C0 −G1 (see Supporting Information Table S1 for details). All .dta files were merged into a single peak list file (.mgf) with an internally developed Perl script. The merged peak list was searched against the human SwissProt database version v57.4 (34 579 sequences, including isoforms as obtained from varsplic.pl) with the search engines MASCOT (v2.2.03, MatrixScience, London, UK) and Phenyx (v2.5.14, GeneBio, Geneva, Switzerland).26 Submission to the search engines was via a Perl script that performs an initial search with relatively broad mass tolerances (MASCOT only) on both the precursor and fragment ions (±10 ppm and ±0.6 Da, respectively). High-confidence peptide identifications were used to recalibrate all precursor and fragment ion masses prior to a second search with narrower mass tolerances (±4 ppm and ±0.025 Da). One missed tryptic cleavage site was allowed. Carbamidomethyl cysteine, N-terminal and lysinemodified iTRAQ 4- or 8-plex was set as a fixed modification, and oxidized methionine was set as a variable modification. To validate the proteins, MASCOT and Phenyx output files were processed by internally developed parsers. Proteins with ≥2 unique peptides above a score T1, or with a single peptide above a score T2, were selected as unambiguous identifications. Additional peptides for these validated proteins with score >T3 were also accepted. For MASCOT and Phenyx, T1, T2 and T3 were equal to 12, 45, 10 and 5.5, 9.5, 3.5, respectively (p-value < 10−3). Following the selection criteria, proteins were grouped on the basis of shared peptides, and only the group reporters are considered in the final output of identified proteins. Spectral conflicts between MASCOT and Phenyx peptide identifications were discarded. The whole procedure was repeated against a reversed database to assess the protein group false discovery rate (FDR). Peptide and protein group identifications were
