Technical Note

The Cell-Surface Proteome of Cultured Adipose Stromal Cells Albert D. Donnenberg,1,2,3* E. Michael Meyer,2 J. Peter Rubin,3,4 Vera S. Donnenberg2,3,5

1

Department of Medicine, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania

2

University of Pittsburgh Cancer Institute, Pittsburgh, Pennsylvania

3

McGowan Institute of Regenerative Medicine, Pittsburgh, Pennsylvania

4

Department of Plastic Surgery, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania

5

Department of Cardiothoracic Surgery, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania

Received 6 January 2015; Revised 18 March 2015; Accepted 9 April 2015 Grant sponsor: National Cancer Institute, Grant number: RO1-CA114246-08 Grant sponsor: University of Pittsburgh Cancer Institute Cytometry Facility, Grant number: CCSG P30CA047904 Additional Supporting Information may be found in the online version of this article. *Correspondence to: Albert D. Donnenberg, Department of Medicine, 5117 Centre Avenue, Suite 2.42 Pittsburgh, Pennsylvania 15213. E-mail: [email protected] Published online 00 Month 2015 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/cyto.a.22682 C 2015 International Society for V

Advancement of Cytometry

Cytometry Part A  00A: 0000, 2015

 Abstract In this technical note we describe a method to evaluate the cell surface proteome of human primary cell cultures and cell lines. The method utilizes the BD Biosciences lyoplate, a system covering 242 surface proteins, glycoproteins, and glycosphingolipids plus relevant isotype controls, automated plate-based flow cytometry, conventional file-level analysis and unsupervised K-means clustering of markers on the basis of percent of positive events and mean fluorescence intensity of positive and total clean events. As an example, we determined the cell surface proteome of cultured adipose stromal cells (ASC) derived from 5 independent clinical isolates. Between-sample agreement of very strongly expressed (n 5 32) and strongly expressed (n =16) markers was excellent, constituting a reliable profile for ASC identification and determination of functional properties. Known mesenchymal markers (CD29, CD44, CD73, CD90, CD105) were among the identified strongly expressed determinants. Among other strongly expressed markers are several that are potentially immunomodulatory including three proteins that protect from complement mediated effects (CD46, CD55, and CD59), two that regulate apoptosis (CD77 and CD95) and several with ectoenzymatic (CD10, CD26, CD13, CD73, and CD143) or receptor tyrosine kinase (CD140b (PDGFR), CD340 (Her-2), EGFR) activity, suggesting mechanisms for the antiinflammatory and tissue remodeling properties of ASC. Because variables are standardized for K-means clustering, results generated using this methodology should be comparable between instrumentation platforms. It is widely generalizable to human primary explant cultures and cells lines and will prove useful to determine how cell passage, culture interventions, and gene expression and silencing affect the cell-surface proteome. VC 2015 International Society for Advancement of Cytometry  Key terms adipose stromal cells; K-means clustering; cell-surface proteomics; lyoplate; highthroughput screening

INTRODUCTION

THE characterization of cell surface protein expression on individual cell types has historically been a gradual process requiring the synthesis of multiple investigations over time. For the best studied cell types such as T lymphocytes, a picture of the cellsurface proteome has gradually emerged, and with it, an in-depth understanding of within-type phenotypic and functional heterogeneity. Often, markers proposed to identify the cell type on which they were first observed (e.g., Thy-1 on thymocytes) have later been demonstrated on disparate cell populations, exposing a misleading nomenclature, which often persists in the literature. Recent advances in high throughput flow cytometry (i.e., plate-based sample loaders), analytical software, and the commercial availability of a product with an array of antibodies directed against human cell surface markers (Becton Dickinson Lyoplate), now facilitate a comprehensive analysis of the cell-surface proteome for culture expanded primary cells and cell lines that lack the intrinsic cellular heterogeneity of tissues. In this report we have determined the cell-surface proteome of cultured adipose stromal

Technical Note cells and provide a method that objectively classifies marker expression into one of four categories on the basis of both percent of positive events and fluorescence intensity. Because standardization of variables is required for this methodology, results should be comparable across platforms.

METHODS Preparation of Adipose Stromal Cells Subcutaneous adipose tissue was harvested during abdominoplasty from five human adult female patients at the University of Pittsburgh Medical Center. All samples were waste materials collected as a byproduct of surgery. Deidentified samples were collected under an IRB-approved exemption (number 0511186, University of Pittsburgh IRB). Upon reception in the laboratory, fat tissue was processed directly for isolation of stromal vascular fraction (SVF), as previously described (1). Fat tissue was minced; digested for 30 min in Hanks’ balanced salt solution (HBSS; Invitrogen), 3.5% bovine serum albumin (BSA; Millipore), and 1 mg/ml collagenase type II (Worthington) on a shaking water bath at 37 8C; and disaggregated through successive 425 and 180 mm sieves (W.S. Tyler). After elimination of mature adipocytes by centrifugation (400 g, ambient temperature, 10 min) and red blood cell lysis (NH4Cl-based buffer [Beckman-Coulter, Cat No. IM3630d], ambient temperature, 10 min), cells were washed in phosphate-buffered saline (PBS) and mononuclear SVF cells were enriched on a Ficoll-Hypaque density gradient (Histopaque-1077; Sigma-Aldrich). Adipose stromal cells (ASC) were selected by adherence after plating SVF at a density of 10–20,000 cells/cm2 in proliferation culture medium (PCM, DMEM, and DMEM/F12, 1:2, 10% fetal bovine serum [FBS] and 0.1 mM dexamethasone [Sigma]) in uncoated T75 or T150 flasks (BD Biosciences Falcon). Culture-expanded cells were trypsinized and carefully cryopreserved in cryoprotectant medium containing 20% newborn calf serum and 10% DMSO. For this study, independent samples were tested at near confluence at passage 3 (3 samples), passage 7, and passage 11. Viability, as measured by Trypan Blue exclusion (Viacell, Beckman Coulter) was 83.4 6 4.6% (mean, SD). Lyoplate Staining Cells were stained for the expression of cell surface markers using a commercially available product, Lyoplate (Becton Dickinson, Cat # 51-9006585AK, Lot 40787). ASC (100,000/well) were stained in the wells according to the manufacturer’s instructions. The Lyoplate Human Cell Surface Marker Screening Panel consists of three 96-well plates, with each well containing a lyophilized antibody to a different cell surface protein (242 human cell surface markers and 9 isotype controls). Antibody binding was detected using R 647–conjugated (AF647) goat antimouse Ig Alexa FluorV and goat antirat Ig detection reagents. To eliminate dead and apoptotic cells, cells were fixed with methanol-free formaldehyde (2% in PBS, 20 min) gently permeabilized with saponin (0.1% in PBS, w/v, 10 min) and stained with DAPI (SigmaAldrich) to a final concentration of 8 mg/ml (2) prior to flow cytometry. 2

Data Acquisition Flow cytometry was performed on an LSRII Fortessa SORP cytometer (BD Biosciences), equipped with a highthroughput sampler (HTS). We acquired 10,323 6 8,885 events (mean 6 SD). The details, including HTS settings, are provided in the Supporting Information MIFlowCyt. Analysis of FCS Files For each of five independent samples, data were exported into FCS files and organized by experiment into playlists using VenturiOne software (Version 6, Applied Cytometry, Dinnington, Sheffield, UK). The gating strategy is shown in Figure 1, where DAPI1 events are selected, gated on forward scatter area and forward scatter width to eliminate cell clusters, and again on DAPI (linear) to select events with 2N DNA content. The marker of interest is then measured on these “clean events” in a histogram of AF647 fluorescence by side scatter. This final analytical gate accounts for slightly higher nonspecific fluorescence in cells with high side scatter. The gate for marker positive events was determined such that all negative controls (isotypes and selected T-cell markers) represented less than 1% of clean events. This “positive gate” was then applied to all wells within an experiment without alteration. Analytical data (event count, percent AF647 positive of clean events, MFI of AF647 positive events, MFI of clean events) were exported to CSV files and imported into Excel, where they were associated with sample id, plate number, plate row, plate column, and reagent. K-Means Clustering Excel data from all 5 samples were imported into SYSTAT (Version 13, SYSTAT Software, San Jose, CA). Markers were considered negative when less than 1% of clean events or less than 100 events were in the positive gate. In such cases MFIpositive data were filled with an arbitrary small number (0.01) plus a random variable (mean 5 0, SD 5 0.01). MFI data were then log10 transformed. Percent positive, log MFIpositive and log MFIclean variables were standardized such that for each variable, the sample mean was subtracted from each value and the difference was divided by the sample standard deviation. The standardized values always have a mean of 0 and a standard deviation of 1. This puts percent and fluorescence intensity data on a common scale allowing them to be evaluated together using multivariate statistical methods. Data from all experiments were partitioned into four clusters together on the basis of standardized percent positive, MFIpositive and MFIclean (Fig. 1, bottom row), using the K-means clustering algorithm (3) as implemented in SYSTAT, using 100 iterations and Euclidean distances. It should be noted that although Kmeans clusters are not inherently ordinal, they were ordered in this implementation (0–3) by inspecting a graph of the original variables grouped by cluster (Fig. 1, bottom). The number of samples assigned to each cluster was then tabulated by marker (shown in Table 2 and in its entirety in Supporting Information). Data were summarized across experiments by marker as median cluster designation [with median absolute deviation (MAD) as a measure of variability], and also mean and standard error of the assigned cluster number. Although these ordered clusters not constitute a continuous variable, Expression of 242 Cell Surface Markers on ASC

Technical Note

Figure 1. Gating strategy and results of K-means clustering. The top 4 rows show the gating strategy used for this analysis and examples of markers classified as clusters 0 (negative), cluster 1 (positive expression on a minor population), cluster 2 (strong expression) through 3 (very strong expression). The bottom row shows the K-means clustering of all markers as a function of standardized percent positive, MFIpositive, and MFIclean. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Cytometry Part A  00A: 0000, 2015

3

4

3

3

CD55

CD59

2 3

CD61 CD90

3

2

CD56

CD46

2 2

CD51/CD61 CD54

2

2 3

CD49d CD49e

CD227

3 3 3

CD49a CD49b CD49C

3

3

CD47

CD166

3 3

CD29 CD44

2

2

CD15

CD164

3

CD9

MARKER

MEDIAN CLUSTER

3.0

3.0

3.0

2.0

3.0

2.0

2.4 3.0

1.6

2.4 2.0

2.2 3.0

3.0 2.6 2.8

3.0

3.0 3.0

2.0

3.0

0.00

0.00

0.00

0.00

0.00

0.00

0.22 0.00

0.54

0.22 0.00

0.18 0.00

0.00 0.22 0.18

0.00

0.00 0.00

0.28

0.00

CLUSTER SEM

Cell adhesion, cell surface signaling Stabilizing cell-cell interactions, endothelial transmigration Cell adhesion, cell surface signaling

Cell adhesion, cell surface signaling Cell-cell and cell matrix interactions, neurite outgrowth, apoptosis, metastasis, inflammation, andfibrosis Cytoprotective anti-adhesive and adhesion receptors, regulation of proliferation Adhesion molecule, Proposed cancel stem cell marker Inhibits cell-cell and cell-extracellula matrix interactions Protection from complement

Integrin a5/lntegrin b3 ICAM-1

Integrin b3 subunit Thy-1

MAC-inhibitory protein

MUC-1, Polymorphic epithelial mucin Complement regulatory protein Complement decayaccelerating factor

ALCAM

Endolyn

Blocks formation of the complement membrane attack complex on the cell surface

Cell adhesion, cell surface signaling Cell adhesion, cell surface signaling

Integrin a4 subunit Integrin a5 subunit

Neural Cell Adhesion Molecule (NCAM)

Apoptosis, proliferation, adhesion, and migration Cell adhesion, cell surface signaling Cell adhesion, cell surface signaling Cell adhesion, cell surface signaling

Adhesion Adhesion, homing, mesenchymal marker

Carbohydrate adhesion molecule

On exosome surface, mediates fusion

BIOLOGICAL PROCESSES

Integrin associated protein Integrin a1 subunit Integrin a2 subunit Integrin a3 subunit

SSEA-1, 3-fucosyl-Nacetyl-lactosamine Integrinb1 Hyaluronic acid receptor

TSPAN-29

ALIAS

MECHANISM(S)

Prevents C9 polymerization

Inactivation of complement components C3b and C4b Interacts with C4b and C3b

Steric hindrance of adhesion interactions

CD6 ligand

Associates with CXCR4

Interacts with fibroblast growth factor receptor, activates p59Fyn signaling pathway Halfofthea5/b3 duplex Upregulatesthrombospondin, SPARC (osteonectin), and fibronectin

Half of integrin duplexes Binding to hyaluronic acid, osteopontin, collagens, and matrix metalloproteinases Thrombospondin-1 receptor, interaction with SIRPa on myeloid cells Half ofthe a1b1 integrin duplex Half of the a2b1 integrin duplex Half of the a3b1 integrin duplex, extracellular matrix receptor Half ofthea4b1 integrin duplex Joins with integrin b 1 to form a fibronectin receptor Recognizes the integrin a5/b3 duplex Binds to CDlla/CD18 or CDllb/CD18

Complexes with integrins and other transmembrane 4 superfamily proteins Expressed on cell surface glycoproteins

Table 1. Markers with very strong (cluster 3) and strong (cluster 2) cell surface expression on cultured ASC

MEAN CLUSTER

Technical Note

Expression of 242 Cell Surface Markers on ASC

MEDIAN CLUSTER

2

3

3 2

3

3

3

2 2

3

3

3

3

3

3

3 3

3

MARKER

CD77

CD95

CD10 CD26

CD13

Cytometry Part A  00A: 0000, 2015

CD73

CD140b

CD142 CD143

CD147

CD2O1

HLAA, B, C

HLAA2

b2 Microglobulin

CD4

CD58 CD63

CD70

1.8

3.0 3.0

1.8

2.8

2.6

3.0

3.0

3.0

2.4 2.0

2.6

3.0

3.0

2.8 2.2

3.0

1.6

MEAN CLUSTER

0.66

0.00 0.00

0.66

0.18

0.22

0.00

0.00

0.00

0.22 0.00

0.22

0.00

0.00

0.18 0.18

0.00

0.22

CLUSTER SEM

CD27 Ligand

LFA-3 LAMP-3

CD4

MHC Class I

MHC Class I A2 serotype

Platelet Derived Growth Factor Receptor B Thromboplastin Hem Progenitor Cell, BB-9, Angiotensin converting enzyme Extracellular matrix metalloproteinase inducer(EMMPRIN) Endothelial protein C receptor Pan MHC Class I

Ecto-50 -nucleotidase

Alanine aminopeptidase

Neprilysin, CALLA Dipeptidyl peptidase-4

Globotriaosylceramide, Gb3 FasR

ALIAS

Presentation of polypeptides to the immune system Presentation of polypeptides to the immune system Presentation of polypeptides to the immune system Expressed on T helper cells, monocytes, macrophages and dendritic cells Present on antigen presenting cells Cell development, activation, growth and motility, cancer metastasis TNF-ligand family cytokine involved in T-cell activation

Enhances protein C serine protease activity

Ectoenzyme, metabolism of regulatory peptides Suppresses T-cell effector responses, mesenchymal marker Recruitment of pericytes and smooth muscle cells to endothelial cells Initiation of thrombin formation Expressed on stromal cells and primitive hematopoietic cells, blood vessel constriction Intracellular recognition in differentiation and development

Peptide hormone inactivation glucose metabolism, fibrosis

Blocks formation of the complement membrane attack complex on the cell surface Cell surface receptor glycosphingolipid involved in apoptosis signaling Programmed cell death receptor

BIOLOGICAL PROCESSES

Table 1. Continued

Co-receptor that assists T-cell receptor/ MHC Class II binding Interacts with CD2 (LFA-2) Signal transduction through integrin complexes Binds to CD27

Interaction with T-cell receptor/CD8 complex Interaction with T-cell receptor/CD8 complex Forms a heterodimer with HLAA,B,C

Receptorfor protein C

Complexes with ubiquitin and transporters

Cell surface tyrosine kinase receptorfor platelet-derived growth factor family Receptor for serine protease factor Vila Converts angiotensin 1 to angiotensin II

Converts AMP to adenosine

Activation of caspase and reactive oxygen species dependent pathways Forms death-inducing signaling complex (DISC) upon Fas ligand binding membrane metallo-endopeptidase Substrates are proline-rich growth factors, chemokines, vasoactive peptides Metalloproteinase

MECHANISM(S)

Technical Note

5

RESULTS

Markers are grouped by function (adhesion, anti-inflammatory, apoptotic, enzymatic, MHC, signal transduction and transport, respectively).

Sodium-independent exchanger 0.00 3 CD98

3.0

c-Cbl Her2/neu, erbB-2 Epidermal growth factor receptor, Herl, erbB-1 Large neutral amino acid transporter (LAT1) 0.00 0.00 0.00 2.0 2.0 2.0 2 2 2 CD165 CD340 EGFR

0.22 0.00 2 3 CD130 CD151

1.6 3.0

IL6-beta Raph Blood Group

Transports branched-chain and aromatic amino acids and other substrates

Signal transduction through JAK/STAT Signal transduction through integrin complexes E3 ubiquitin-protein ligase Receptor tyrosine kinase Receptor tyrosine kinase

TGFb auxiliary receptor Activation of the protein kinase PTK2/FAK1 0.00 0.18 CD105 CD108

3.0 2.8 3 3

CD81

0.00 3

MARKER

CLUSTER SEM MEDIAN CLUSTER

3.0

Endoglin Semaphorin 7A

BIOLOGICAL PROCESSES

Regulation of cell development, activation, growth and motility. Candidate tumor suppressor Neoangiogenesis, Fibrosis Regulation of cell migration, focal adhesion complexes Type-1 cytokine receptor subunit Activation, growth, motility, MMP induction Cell signaling, protein ubiquitination Signal transduction in multiple pathways Signal transduction in multiple pathways

MECHANISM(S)

6

parametric statistics (mean, SD) were far more informative than nonparametric statistics, as the MAD was 0, and the mode was equal to the median in almost every case.

MEAN CLUSTER

TAPA-1

ALIAS

Table 1. Continued

Interacts with TSPAN4, CD19,CD9,PTGFRN,CD 117andCD29

Technical Note

Interpretation of Clusters The decision to classify the markers into four clusters on the basis of percent positive events, mean fluorescence intensity of positive events, and mean fluorescence intensity of all clean events was made empirically and yielded four distinct clusters (Fig. 1). It should be noted that K-means clusters are inherently nonordinal, but our goal was to quantify the expression of each marker in terms of a metric including both fluorescence intensity and percent of positive events. To accomplish this, we plotted the standardized variables by cluster (Fig. 1, bottom), and reordered the cluster numbers such that 0 represented the cluster with the negative markers and 3 represented the cluster containing the markers with the highest percent positive. Markers partitioning into cluster 3 were homogenous for percent positive (97.1 6 3.5%, mean, SD), with positive events having high fluorescence intensity (geometric mean MFIpositive 5 25,874, 5th and 95th percentiles (CI) 5 8,682 2 109,412). Cluster 2 averaged 64.8 6 16.4% positive, with an average MFIpositive of 4,914, CI 5 2,440–18,091). Markers clustering into Group 1 were interesting because some revealed intersample heterogeneity with clear positive and negative populations (e.g., CD36, Fig. 1), whereas others were equivocal. Cluster 1 markers had a lower mean percent positive than cluster 2 (8.2 6 7.6%), but a slightly higher geometric mean MFIpositive (5,415, CI 5 2,276–25,283). Markers partitioning into cluster 0 were uniformly negative (0.4 6 0.8% positive). Between-sample agreement was striking, with most markers having a median absolute deviation of 0. For this reason, the mean and standard error of cluster number were also used as summary statistics, because high standard errors (>0.22) identified the few markers that were discordant between samples. Negative Controls All isotype (mouse IgG1, IgGa, IgG1b, Ig3, IgM, rat IgG1, IgG2a, IgG2b, IgM) and antiglobulin markers (AF647 goat anti-mouse, AF647 goat anti-rat) clustered in Group 0 in all cases, as did T-cell markers CD2, CD3, alpha-beta and gamma-delta TCR, and B-cell markers CD19 and CD20. In all, 171 markers partitioned into cluster 0 (median cluster 5 0, MAD 5 0, mean cluster < 0.6, SEM  0.22). Very Strong (Cluster 3) and Strong (Cluster 2) Positive Markers Between-sample agreement in very strong and strong markers was striking (Table 1). A total of 32 markers clustered as very strong (median cluster 5 3, MAD 5 0, mean cluster  2.6, SEM  0.22) and 16 clustered as strong (median cluster 5 2, MAD 5 0, mean cluster  1.6 and

The cell-surface proteome of cultured adipose stromal cells.

In this technical note we describe a method to evaluate the cell surface proteome of human primary cell cultures and cell lines. The method utilizes t...
333KB Sizes 4 Downloads 11 Views