| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Institute of Pharmaceutical Sciences, Swiss Federal Institute of Technology, ETH Zurich, Zurich, Switzerland
2Correspondence: Institute of Pharmaceutical Sciences, Swiss Federal Institute of Technology, ETH Zurich, Wolfgang-Pauli-Str. 10, HCI H303, CH-8093 Zurich, Switzerland. E-mail: michael.detmar{at}pharma.ethz.ch
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: angiogenesis • lymphangiogenesis • endothelium • proteomics • membrane proteins
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3)
, there has been a quest for identifying the molecular mechanisms that control lymphatic and blood vessel formation and function in development and disease (4)
.
One of the obstacles to studying angiogenesis and lymphangiogenesis is the limited availability of specific markers for the cell types involved in these processes. Recent studies (1
, 5
, 6)
in genetically engineered mouse models have identified several key molecules that govern the embryonic formation of the blood vasculature and the consecutive development of the lymphatic system from embryonic veins. Moreover, it has recently become possible to isolate and culture blood vascular endothelial cells (BECs) and lymphatic endothelial cells (LECs) from human tissues. These cells maintain their lineage-specific differentiation even after several passages in vitro (7
8
9)
, enabling functional and molecular investigations of their cell type-specific properties. Recent transcriptional profiling studies (7
, 10
11
12
13
14)
have indeed revealed several previously unknown lineage-specific markers for blood vascular and lymphatic endothelium and have led to the identification of novel lymphangiogenic pathways. Previous proteomic studies (15
16
17
18
19)
have selectively investigated the proteome of resting or activated blood vascular endothelium in vitro and in vivo. However, a comprehensive evaluation of vascular lineage-specific protein expression has been lacking.
Because surface-accessible proteins might serve as suitable targets for imaging and/or inhibiting pathological angiogenesis (20)
and lymphangiogenesis associated with cancer metastasis and other diseases, we aimed to identify the surface-accessible, lineage-specific vascular proteome. Membrane proteins represent up to two-thirds of all drug targets (21)
, yet these proteins often escape proteomic analysis by 2D-PAGE because of their low abundance and their poor solubility in aqueous buffers (15
, 22)
. To circumvent this problem, we applied a gel-free proteomic methodology termed two-dimensional peptide mapping (2D-PM) (15
, 23)
. This technique allows for a comprehensive identification of surface-accessible proteins.
For 2D-PM, membrane and extracellular matrix proteins are labeled with a cleavable biotin reagent. The labeled proteins are then isolated using streptavidin beads and digested with trypsin, and the resulting peptides are separated by reverse-phase HPLC. Arising fractions are mixed with internal standard peptides and automatically spotted onto matrix-assisted laser desorption/ionization (MALDI)-target plates. The data obtained by MALDI-time-of-flight (TOF) mass spectrometry (MS) are then used to establish 2D peptide maps that display the HPLC fractions, the mass-to-charge (m/z) ratios, and the normalized, relative signal intensities of the measured peptides. We used the 2D-PM technique with human BECs and LECs, stimulated by incubation with vascular endothelial growth factor (VEGF)-A, to create reproducible and quantitative 2D peptide maps of lineage-specific proteins and also to identify new lineage-specific vascular markers that might serve as potential new targets for the selective imaging and/or targeting of lymphatic and blood vessels. The simplicity of this method as well as the ability to comparatively quantify protein expression establish 2D-PM as a powerful tool for large-scale proteomic analyses of membrane and extracellular matrix proteins and might also simplify the discovery of biomarkers and drug targets in other model systems.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 11)
and were propagated on collagen-coated (50 µg/ml; Angiotech BioMaterials Corp., Palo Alto, CA, USA) cell culture flasks for up to eight passages. All LEC cultures expressed high levels of mRNAs encoding the major lymphatic lineage markers Prox1 (a homeobox transcription factor) and LYVE-1 (a hyaluronan receptor). These markers were either not expressed or expressed at very low levels by BECs, which expressed instead the blood vascular lineage marker Flt-1. Cells were cultured in endothelial cell basal medium (Cambrex, East Rutherford, NJ, USA), supplemented with 10 µg/ml hydrocortisone acetate, 25 µg/ml N-6,2'-O-dibutyryl-adenosine 3',5'-cyclic monophosphate, antibiotic-antimycotic solution, 2 mmol/L L-glutamine (all from Sigma, Buchs, Switzerland), 20% fetal bovine serum (Life Technologies, Inc., Invitrogen, Carlsbad, CA, USA), and 20 ng/ml recombinant human VEGF-A165 (NCI Biological Resources Branch, Frederick, MD, USA). VEGF-A was continuously added as a culture supplement to all cultures because it was needed to efficiently expand the BEC and LEC cultures for up to eight passages and because we were interested in proteins that might also be expressed on the surface of activated vessels. For each cell type, twelve 300 cm2 cell culture flasks containing confluent monolayers of LECs or BECs at passage 8 (
6x106 cells per flask) were used for cell surface biotinylation as described previously (15)
. Three independent experiments were performed. All solutions used for biotinylation were cooled to 4°C. Cells were washed once with 30 ml of PBS to remove nonviable cells, followed by incubation with 15 ml of PBS containing 1.6 µM EZ-link sulfo-NHS-SS-biotin (Pierce, Rockford, IL, USA) for 5 min at room temperature on a shaker. During this process, the biotin-tag is covalently linked to primary amines of accessible proteins expressed on the cell surface. To terminate the biotinylation reaction, 1 M Trizma Base (Fluka, Buchs, Switzerland) in H2O was added to a final concentration of 16 µM. Cells were detached by scraping into 10 ml of a 10 mM EDTA/PBS solution containing 1.6 µM oxidized glutathione (Fluka) to prevent reduction of the disulfide bond. After centrifugation, cell pellets were washed once with 20 ml of PBS containing 1.6 µM oxidized glutathione. Cell viability was evaluated by staining of trypsinized LECs and BECs at passage 8 with trypan blue (Sigma), followed by quantitative analysis of dead and viable cells.
Isolation, elution, and digestion of biotinylated proteins
For each 300 cm2 cell culture flask, cells were lysed for 30 min on ice with 1 ml of lysis buffer [2% wt/vol Nonidet P-40 substitute (Fluka), 0.2% wt/vol sodium dodecyl sulfate (SDS), 1x complete E protease inhibitor (Roche Diagnostics, Rotkreuz, Switzerland), 10 mM EDTA, and 108 µM oxidized glutathione in PBS]. The cell lysate was centrifuged for 10 min at 16,100 g, and the cleared supernatant was used for purification of biotinylated proteins on streptavidin-Sepharose high performance (GE Healthcare, Piscataway, NJ, USA). Three individual samples of BECs (B1, B2, and B3) and of LECs (L1, L2, and L3) were prepared, containing 3 mg of total protein each (determined by BCA protein assay; Pierce). Then, 640 µl of streptavidin-Sepharose slurry was washed 3x with 500 µl of washing buffer A (1% wt/vol Nonidet P-40 substitute, 0.1% wt/vol SDS, and 20 mM oxidized glutathione in PBS) before the cleared lysates were added. The total SDS concentration was adjusted to 2% wt/vol. The samples were tumbled for 2 h at room temperature before removal of unbound proteins by washing 3x with buffer A, 2x with buffer B (2 M NaCl, 0.1% wt/vol SDS, and 20 mM oxidized glutathione in PBS), and 2x with 50 mM Tris-HCl, pH 7.5. Captured proteins were eluted from the streptavidin-Sepharose by incubation with 400 µl of 5% 2-mercaptoethanol in PBS for 30 min at 30°C. The elution step was repeated 3x; eluates were pooled and split. Samples were precipitated by addition of 100 µl of 100% wt/vol trichloroacetic acid and incubated for 30 min on ice. The precipitates were pelleted by centrifugation (16,100 g for 5 min), washed with ethanol-ether (1:1), and air-dried. Each precipitate was dissolved in 200 µl of digestion buffer (50 mM Tris-HCl, pH 8.0, and 1 mM CaCl2), and 1.6 µg of sequencing grade modified trypsin (Promega, Madison, WI, USA) was added. The digestion was carried out overnight at 37°C in a thermomixer (Eppendorf, Hamburg, Germany) at 1000 rpm. Resulting tryptic peptides were desalted, purified, and concentrated with C18 microcolumns (ZipTip C18, Millipore, Billerica, MA, USA) according to the manufacturer’s instructions. Samples were lyophilized and stored at –20°C.
Reverse phase-HPLC
Trypsin-cleaved peptides were separated by reverse phase HPLC using an UltiMate nanoscale LC system and a FAMOS microautosampler (Dionex, Sunnyvale, CA, USA) controlled by the Chromeleon software (Dionex). Mobile phase A consisted of 2% acetonitrile and 0.1% trifluoroacetic acid (TFA) in water, and mobile phase B consisted of 80% acetonitrile and 0.1% TFA in water. The flow rate was set to 300 nl/min. Each of the six samples (B1-B3 and L1-L3) was dissolved in 12 µl buffer A, and 5 µl was loaded on the column (15 cmx75 µm inner diameter, C18 PepMap 100, 3 µm, and 100 Å; Dionex). The peptides were eluted with a gradient of 0% buffer B for 3 min, 0–60% buffer B for 81 min, 60–100% buffer B for 10 min, and 100% buffer B for 5 min; the column was equilibrated with 100% buffer A for 20 min before the next sample was analyzed. The eluted fractions were mixed with a solution of 3 mg/ml 
-cyano-4-hydroxycinnamic acid, 277 pmol/ml of each of the three internal standard peptides (neurotensin, angiotensin I, and adrenocorticotropic hormone fragment 1–17; all from Sigma), 70% acetonitrile, and 0.1% TFA in water and were deposited on a 192-well MALDI target plate using the on-line Probot system (Dionex). The flow of the MALDI matrix solution was set to 1083 nl/min. Thus, fractions collected each 20 s contained 361 nl MALDI-matrix solution and 100 nl of sample. The final concentration of each of the three standard peptides was 100 fmol per spot.
MALDI-TOF/TOF
MALDI-TOF and MALDI-TOF/TOF mass spectrometric analyses were carried out with a 4700 Proteomics Analyzer (Applied Biosystems, Foster City, CA, USA). Peptide masses were acquired over a range from 750 to 4000 m/z, with a focus mass of 2000 m/z. MS spectra were summed from 2000 laser shots from an Nd:YAG laser operating at 355 nm and 200 Hz. An automated plate calibration was performed using five peptide standards (masses 900-2400 m/z; Applied Biosystems) in six calibration wells. This plate calibration was used to update the instrument default mass calibration, which was applied to all MS and MS/MS spectra. Furthermore, an internal calibration of each MS spectrum using the three internal standard peptides added to the MALDI matrix was performed. A maximum of 10 precursors per sample well with a signal-to-noise ratio of >100 was automatically selected for subsequent fragmentation by collision induced dissociation. MS/MS spectra were summed from 2500 to 5000 laser shots. Spectra were processed and analyzed by the Global Protein Server Workstation (Applied Biosystems), which uses internal MASCOT (Matrix Science, London, UK) software for matching MS and MS/MS data against databases of in silico digested proteins. The MASCOT search parameters were 1) a human database downloaded from the European Bioinformatics Institute homepage (http://ftp.ebi.ac.uk/pub/databases/SPproteomes/fasta/proteomes/25.H_sapiens.fasta.gz); 2) allowed number of missed cleavages: one; 3) variable post-translational modifications: methionine oxidation; 4) peptide tolerance: ±10 ppm; 5) MS/MS tolerance: ±0.2 Da; and 6) peptide charge: +1. Peptides were selected for analysis if identified with a confidence interval of >95%, using the Mowse scoring algorithm, which calculates protein identification probability (24)
. Nomenclature was determined based on Swiss-Prot primary accession numbers according to the UniProtKB/Swiss-Prot entry on 13 July 2006. Redundancies in protein identification were removed.
2D-PM
The MALDI-TOF MS data were used to establish 2D peptide maps with the Spectational software (15)
, which plots HPLC fractions on the y axis and the m/z ratios of the measured peptides on the x axis. In addition, the software translates the logarithmic values of the normalized intensities to a gray scale. The normalization is performed with respect to the internal standard peptide neurotensin. ASCII spectrum files for the creation of 2D peptide maps were produced using Data Explorer V4.6 software (Applied Biosystems). The same software was used to plot the MALDI-TOF MS spectra shown in
Fig. 2
.
|
|
Immunoblotting
LECs and BECs were lysed as described above. For each sample, 25 µg of protein was separated by denaturing SDS-PAGE on 4–12% Bis-Tris gels (Invitrogen) and the protein was then transferred to nitrocellulose membranes (Whatman, Brentford, UK), followed by 1 h blockage in PBS containing 5% defatted milk. The following antibodies were diluted in PBS containing 5% defatted milk and were incubated with the membranes for 1 h: mouse anti-beta-actin (Sigma); goat anti-aldehyde dehydrogenase 1A1, rabbit anti-fibronectin, and goat anti-vascular endothelial (VE)-cadherin (all from Santa Cruz Biotechnology, Santa Cruz, CA); mouse anti-CD31 (DakoCytomation, Glostrup, Denmark); mouse anti-drebrin (Progen, Heidelberg, Germany); mouse anti-endoglin (NeoMarkers, Thermo Fisher Scientific Inc., Fremont, CA); mouse anti-LASP-1 (Chemicon International, Millipore, Billerica, MA, USA); and mouse anti-N-cadherin (BD Biosciences, San Jose, CA, USA). To detect angiopoietin-2 and jagged-1, 50 µg of each protein sample was used in the SDS-PAGE and goat anti-angiopoietin-2 (Santa Cruz Biotechnology) and goat anti-jagged-1 (R&D Systems, Minneapolis, MN, USA) antibodies were used for immunoblotting. To detect multimerin-1 (MMRN1), 100 µg of each protein sample and NuPAGE antioxidant (Invitrogen) was used in the SDS-PAGE and rabbit-anti MMRN1 (kindly provided by C. P. M. Hayward, McMaster University, Hamilton, ON, Canada) was used for immunblotting. To detect ESAM, 40 µg of each protein sample was separated by 10% Bis-Tris SDS-PAGE, transferred to nitrocellulose membranes (Bio-Rad, Hercules, CA, USA), and immunoblotting was performed with a goat anti-endothelial cell-selective adhesion molecule (ESAM) antibody (R&D Systems). Donkey anti-rabbit-IgG-HRP (GE Healthcare), sheep anti-mouse-HRP (Amersham, GE Healthcare), and rabbit anti-goat-HRP (Sigma) were used as secondary antibodies. Antibody reactions were visualized using the Amersham ECL plus Western blotting detection system (GE Healthcare). The loading control was performed by incubation with SYPRO Ruby (Molecular Probes, Invitrogen) for 3 h. After washes in distilled water overnight, the gels were imaged with a DIANA III fluorescence imager (Raytest, Straubenhardt, Germany).
Immunohistochemistry
Immunofluorescence analysis was performed on 6 µm cryostat sections of neonatal human foreskins as described previously (7)
, using antibodies against retinal dehydrogenase 1 (ALDH1A1), ESAM, MMRN1 (all as described above), CD31 (DakoCytomation), CD34 (BD Pharmingen), and podoplanin (clone D2–40; Signet Laboratories, Dedham, MA, USA). Immunofluorescence was also performed on 6 µm cryosections of mouse skin, small intestine, and lung and of xenotransplants of MDA-435 cells stably transfected with VEGF-C (25)
, using a goat anti-ESAM antibody (R&D Systems; ref. 26
), rat Meca32 antibody (specific for blood vessels; BD Pharmingen), and a rabbit-anti LYVE-1 antibody (specific for lymphatic vessels; Angiobio, Del Mar, CA, USA). The corresponding secondary antibodies were labeled with AlexaFluor488 or AlexaFluor594 (Molecular Probes). Nuclei were counterstained with 100 ng/ml Hoechst 33342 (Molecular Probes). Sections were examined using an Axioskop2 mot plus microscope (Zeiss, Oberkochen, Germany), and images were captured with an AxioCam MRc camera (Zeiss) using the AxioVision 4.4 software at x400. Confocal laser scanning microscopy was performed on a LSM510 Meta laser scanning microscope (Zeiss), and images were analyzed using the Imaris Software (Bitplane, Zurich, Switzerland).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Identification of vascular lineage-specific proteins by 2D-PM
All 192 MALDI-TOF spectra were used to create 2D peptide maps of each sample using the Spectational software package (15)
. Representative maps from BEC and LEC samples are shown in Fig. 1
. In each 2D peptide map, the m/z ratio was plotted for each HPLC fraction and neurotensin was used as an internal standard. Each 2D peptide map contained several hundred spots (Fig. 1)
with each spot corresponding to a distinct peptide. Most peptides eluted in more than one HPLC fraction.
Analysis of the individual MALDI-TOF/TOF spectra using the MASCOT software, to match MS and tandem mass (MS/MS) data against databases of in silico digested proteins, resulted in an average of 350 identified proteins for each individual sample. Two typical MS and MS/MS spectra and a list of identified fragment ions are shown in Fig. 2
. In total, 746 different proteins were identified in all six samples of BECs and LECs (Supplemental Table 1); 184 proteins were specifically or predominantly expressed by LECs, 185 proteins were specifically expressed by BECs, and 377 proteins were comparatively expressed by both cell types (common endothelial proteins). Importantly, many of the common endothelial proteins were detected in all 6 samples, and up to 20 different peptides were identified for each distinct protein (Supplemental Table 1). The most abundantly expressed common endothelial proteins were the platelet-endothelial cell adhesion molecule-1 (PECAM-1; CD31), VE-cadherin, and the integrin alpha-2 subunit (Table 1
; Supplemental Table 2): these proteins were previously known to be expressed by both types of endothelial cells (27)
. We also identified several proteins that had not been previously associated with BECs or LECs, such as cadherin-23 (Supplemental Table 1).
|
In addition to previously described markers of the lymphatic endothelium such as carcinoembryonic antigen-related cell adhesion molecule 1 (CEACAM1) (10)
, 2D-PM also identified a number of new proteins that were specifically expressed by LECs. The most abundant of these was ALDH1A1: a total of 16 different ALDH1A1 peptide fragments was detected in all three LEC samples, whereas no ALDH1A1 peptide was detected in the BEC samples. Other LEC-specific proteins included MMRN1 and LIM and SH3 domain protein 1 (LASP-1). Among the BEC-specific proteins, we identified several proteins that were not previously known to be restricted to or to be more strongly expressed by blood vascular endothelium including hedgehog-interacting protein (HHIP), CYR61 protein and drebrin, and acting-binding protein. Several known markers of BECs were also detected by the assay, including neuronal cell adhesion molecule (NCAM), N-cadherin, and neuropilin-1 (7
, 12
; Supplemental Table 1).
Detailed analysis of select marker proteins
Twenty-seven proteins were selected for a more detailed analysis (9 representative proteins per group of common endothelial, LEC-specific, or BEC-specific proteins; Table 1
; Supplemental Table 2). The majority of the selected proteins was annotated as single-pass type I membrane proteins (52%) or extracellular matrix (ECM) proteins (22%). The remaining proteins were annotated as cytoplasmic, shuttling between the cytoplasm and plasma membrane, secreted, and multipass membrane proteins. After normalization of the individual MS spectra, we confirmed that for all nine common endothelial proteins evaluated, peptides corresponding to these proteins showed similar expression levels in both BECs and LECs (Fig. 3
A). For the nine proteins that were specifically associated with BECs, the normalized signal intensity was much higher in BECs than in LECs (Fig. 3b
). Conversely, increased normalized signal intensities were observed for all nine selected LEC-specific proteins (Fig. 3C
). For some proteins identified only in the LEC samples, no corresponding peptide peak could be found in the BEC samples, indicating exclusive expression of these proteins in LECs.
|
Validation of markers by immunoblot analysis
Out of the 27 selected marker proteins, 13 were chosen for validation by immunoblot analysis. We found that the common endothelial marker proteins PECAM-1, VE-cadherin, and endoglin were expressed at comparable levels in both LECs and BECs (Fig. 4
A). Fibronectin, beta-catenin, drebrin, N-cadherin, and ESAM were more highly expressed or exclusively expressed in BECs, compared with LECs (Fig. 4B
), whereas ALDH1A1, angiopoietin-2, Jagged-1, LASP-1, and MMRN1 were more strongly or exclusively expressed by LECs (Fig. 4C
). The signals of the beta-actin immunoblot (Fig. 4D
) and of the SYPRO Ruby staining (Supplemental Fig. 1) were comparable for both cell types, confirming loading of equal amounts of proteins. Thus, immunoblot analyses confirmed the differential expression patterns of vascular markers identified by 2D-PM.
|
In situ validation of markers by immunofluorescence analysis of human and mouse tissue samples
To further investigate whether the lineage-specific marker proteins identified by 2D-PM in vitro also showed vascular lineage-specific expression in situ, we performed differential immunofluorescence analysis of human foreskin sections. To this end, we used antibodies against the lymphatic marker proteins ALDH1A1 and MMRN1 and the blood vascular marker protein ESAM identified by 2D-PM, combined with antibodies against the blood vascular marker CD34, the lymphatic vessel marker podoplanin, or the pan-endothelial marker CD31. We found that ESAM was expressed by CD34-positive blood vessels, in which it had a membrane staining pattern but not by podoplanin-positive lymphatic vessels (Fig. 5
Aa–f). Conversely, an antibody against ALDH1A1 selectively stained the membranes of podoplanin-positive lymphatic vessels but not of CD34-positive blood vessels (Fig. 5Ag-l
). ALDH1A1 staining was pronounced at the membrane of D2–40 positive lymphatic vessels in human foreskin sections as shown by confocal laser scanning microscopy (Fig. 5Ba-f
). MMRN1 was expressed by D2–40 positive lymphatic vessels but only by a fraction of the CD31-positive vessels and not by CD34-positive blood vessels (Fig. 5Am-u
).
|
We next investigated the expression of ESAM in normal mouse tissues and in tumor-associated vessels in xenotransplants of MDA-435 cells expressing VEGF-C, together with antibodies against the blood vascular marker Meca32 and the lymphatic vessel marker LYVE-1. ESAM was expressed by Meca32-positive blood vessels but not by LYVE-1-positive lymphatic vessels in mouse skin, small intestine (Fig. 6
a–p), lung (data not shown), and tumor xenotransplants (Fig. 6q-x
).
|
Taken together, these results confirm the lineage-specific expression of selected markers, identified by 2D-PM of cultured cells, in the vasculature of human and mouse tissue samples.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 28
29
30
31
32)
. Moreover, chronic inflammatory diseases are characterized by pronounced involvement of both blood and lymphatic vessels (2
, 5
, 27
, 33)
. Specific markers of the formation of new blood or lymphatic vessels can therefore not only be developed as therapeutic targets but can also be used in imaging and the development of diagnostics for cancer and inflammatory diseases (5
, 27)
.
Because vascular targeting and imaging largely rely on the specific binding of compounds to proteins that are accessible from the bloodstream or from the lymphatic fluid (20
, 31)
, we aimed to identify membrane and ECM proteins expressed by blood vascular and/or lymphatic endothelial cells. The technology we applied to do so, 2D-PM, differs from conventional 2D gel-based techniques in that it is compatible with the strong anionic detergents necessary for the efficient solubilization of most membrane proteins and ECM components. The use of a cleavable, charged biotin derivative increases the yield of surface-accessible proteins, allowing the affinity purification and enrichment of labeled proteins by streptavidin-Sepharose. The 2D-PM methodology also produces peptide maps that can be easily analyzed for the comparative expression levels of a large number of distinct proteins.
Using 2D-PM, we identified a total of 746 proteins, 377 of which were expressed equally by BECs and LECs. These proteins included well known pan-endothelial markers such as PECAM-1 and VE-cadherin, which validates this methodology, as well as a number of new vascular markers, which were validated independently by immunoblotting and immunofluorescence analysis of human tissue samples. Although a number of proteins were identified in the mass spectrometric analysis with only one peptide, we have used stringent identification criteria (95% confidence interval for each peptide) and have confirmed the differential regulation of selected of these proteins by immunoblot analyses (e.g., angiopoietin-2 and Jagged-1). The protein expression patterns were also reproducibly confirmed with replicate samples of BECs and LECs. The 2D-PM technology enables the reliable, comparative quantification of protein expression between different biological samples. The comparison of expression levels can be achieved by three complementary analyses: the number of replicates in which peptide fragments of a specific protein are detected, the number of distinct tryptic peptides detected in all samples of an experimental group, and most importantly the direct comparison of normalized signal intensities of the same peptide.
Although our 2D-PM analysis yielded a strong enrichment of membrane and extracellular matrix proteins, there were also several proteins detected that are annotated as intracellular proteins. This finding could be due to an incorrect or incomplete annotation of some of the proteins. For example, nucleolin is annotated as a nuclear protein but becomes accessible on the membrane of proliferating endothelial cells (34)
. In the present study, we have shown by confocal laser scanning microscopy that ALDH1A1 is localized on the cell membranes of lymphatic vessels in human skin samples, although it is annotated as a cytosolic protein. Moreover, in a recent proteomic study in F9 tumor cells (unpublished results), we observed the membrane localization of alpha-actinin-4, high mobility group protein B1, and adenylate kinase 2. All these proteins have previously been described to be intracellular. In addition, the biotin reagent might have gained access to the intracellular compartment of a few nonviable cells. Because the total number of intracellular proteins is higher than that of membrane proteins, and because intracellular proteins often have higher copy numbers than membrane proteins, a contamination by nonviable cells may result in the identification of intracellular proteins. However, viability tests revealed that the number of nonviable cells was below 3% before addition of the biotin reagent.
It is of interest that several known markers for LECs or BECs, including podoplanin and LYVE-1, were not identified in our study. Since the biotinylation reagent interacts with primary amine groups (lysine side chain or protein N terminus), a lack of primary amine in the extracellular domain of membrane spanning proteins, as well as an inaccessibility due to the three-dimensional structure of the protein, may prevent efficient biotinylation. Moreover, insufficient fragmentation of distinct peptides during MS/MS may prevent their identification by the currently used software tools, and post-translational modifications further complicate the identification of peptides and proteins.
Transcriptomic analyses of cultured LECs and BECs have previously identified a number of lineage-specific genes (7
, 12)
and have also led to the identification of novel lymphangiogenic pathways (11)
. Our proteomic analysis confirmed the lineage specificity of several of the BEC- and/or LEC-specific genes identified by transcriptional profiling, including fibronectin, beta-catenin, NCAM, N-cadherin, angiopoietin-2, jagged-1, and CEACAM (7
, 10
, 12)
. Moreover, 2D-PM also identified a number of previously unidentified lymphatic or blood vascular lineage-specific proteins. Differences between previous transcriptomic and our proteomic study are likely due to post-transcriptional and post-translational modification processes, and they further underline the importance of investigations at the level of protein expression.
The vessel-associated proteins identified in the 2D-PM analysis also provide some new insights into the mechanisms of angiogenesis and lymphangiogenesis. The HHIP produced the largest number of BEC-associated peptides. HHIP is an antagonist of hedgehog signaling, and previous studies (35)
showed that HHIP mRNA was expressed by vascular endothelial cells. Our proteomic data, showing that HHIP is expressed in the blood vascular but not the lymphatic endothelium, indicate a possible role for the hedgehog pathway in vascular lineage identity. Moreover, we identified drebrin, a protein that links actin filaments to membrane proteins, as a BEC-specific protein. Drebrin was first detected in neuronal cells but has recently been found to also be expressed in human skin (36)
. The results of our study indicate that drebrin is another important protein in the lineage-specific function of blood vascular endothelium.
The protein with the highest LEC-specific peptide count was ALDH1A1, which was also shown by immunoblotting to be specifically expressed by LEC. ALDH1A1 regulates the formation of retinoic acid, a signaling molecule that controls cellular differentiation during embryonic development (37)
. Based on the potent effects of retinoids on cell differentiation, retinoid signaling might also be involved in cell fate decisions during lymphatic vascular development. MMRN1 was identified as a lymphatic-specific protein by 2D-PM. Previous studies of MMRN1 reported storage in platelet alpha granules and in endothelial Weibel-Palade bodies, where it should not be accessible by surface biotinylation. Thus, release of MMRN1 from Weibel-Palade bodies in VEGF-A-activated LECs likely rendered the protein accessible for biotinylation in our study.
Taken together, these results reveal 2D-PM as a powerful technology to identify lineage-specific, surface-accessible proteins and to compare the relative expression levels of these proteins between different cell types. This methodology is robust and reproducible, and the differential expression of blood and lymphatic vessel proteins was confirmed by immunoblotting in vitro and by immunohistochemistry in situ. The proteins most strongly associated with each cell type, HHIP (BECs) and ALDH1A1 (LECs), might serve as novel markers for these cell lineages. Further studies are needed to evaluate which of the newly identified lineage-specific proteins might represent the most promising targets for therapeutic and/or imaging strategies. This novel proteomic approach for the identification of surface-accessible proteins might also simplify the discovery of biomarkers and drug targets in other model systems.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication October 15, 2007. Accepted for publication December 6, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  B. Borgia, C. Roesli, T. Fugmann, C. Schliemann, M. Cesca, D. Neri, and R. Giavazzi A Proteomic Approach for the Identification of Vascular Markers of Liver Metastasis Cancer Res., January 1, 2010; 70(1): 309 - 318. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. E. Kalin, N. E. Banziger-Tobler, M. Detmar, and A. W. Brandli An in vivo chemical library screen in Xenopus tadpoles reveals novel pathways involved in angiogenesis and lymphangiogenesis Blood, July 30, 2009; 114(5): 1110 - 1122. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Roesli, B. Borgia, C. Schliemann, M. Gunthert, H. Wunderli-Allenspach, R. Giavazzi, and D. Neri Comparative Analysis of the Membrane Proteome of Closely Related Metastatic and Nonmetastatic Tumor Cells Cancer Res., July 1, 2009; 69(13): 5406 - 5414. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
