Rapid Comparative Immunophenotyping of ... - Wiley Online Library

Download PDF
2 downloads 0 Views 814KB Size Report
Comment
Abstract. Flow cytometry immunophenotyping is a sensitive technique allowing rapid characteri- zation of single cells within heterogeneous populations, but it ...
Original Article

Rapid Comparative Immunophenotyping of Human Mesenchymal Stromal Cells by a Modified Fluorescent Cell Barcoding Flow Cytometric Assay Tamara Lekishvili,* Jonathan J. Campbell

LGC, Queens Road, Teddington, Middlesex, TW11 0LY, United Kingdom Received 22 February 2017; Revised 30 June 2017; Accepted 2 September 2017 Grant sponsor: UK government Department for Business, Energy & Industrial Strategy (BEIS). Additional Supporting Information may be found in the online version of this article. *Correspondence to Dr Tamara Lekishvili, LGC, Queens Road, Teddington, Middlesex, TW11 0LY, United Kingdom. Email: [email protected] Published online 00 Month 2017 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/cyto.a.23248 C 2017 International Society for V

Advancement of Cytometry

 Abstract Flow cytometry immunophenotyping is a sensitive technique allowing rapid characterization of single cells within heterogeneous populations, but it includes several subjective steps during sample analysis that impact the development of standardized methodology. Proposed strategies to overcome these limitations include fluorescent cell barcoding (FCB), which facilitates multiplexed sample evaluation with increased data reproducibility whilst reducing labeling variation, materials, and time. To date, the FCB assay has found utility for analyzing the phosphorylation status of intracellular targets but has not been intensively employed for cellular immunophenotypic analyses using cell surface markers. In this study we developed a modified FCB assay for multiplexed analysis of human mesenchymal stromal cells (hMSCs) to evaluate the quality of these cells during bioprocessing. A panel of fluorochrome-conjugated antibodies was used to target 15 ubiquitously expressed or stage-specific markers together with a fixable viability dye eFluor 506 acting as the cell barcoding agent. Critical technical considerations and validation steps were presented in the context of monitoring hMSC status, defined by generic, and specific surface markers for cell identity and quality. It was found that at discrete passages, inter-analyst expression patterns between hMSCs cultures were similar, but in contrast, diverse marker expression was evident between passages. A side-by-side analysis of barcoded and non-barcoded cells demonstrated the potential of this technique for the rapid phenotypic characterization of cells exposed to different bioprocessing conditions. Additionally, the method incorporates fewer subjective factors; including sample preparation and instrument day-to-day variations and is customizable across a diversity of cell types. VC 2017 International Society for Advancement of Cytometry

 Key terms flow cytometry; assay standardization; fluorescent cell barcoding; immunophenotyping; fixable viability dye; human mesenchymal stromal cells

HUMAN

mesenchymal stromal cells (hMSCs) remain promising candidates for tissue engineering and regenerative medicine applications due to their tissue repair potential and immunomodulatory function (1–4). However, cellular phenotype can vary widely within the initial set of criteria already established by the International Society for Cellular Therapy (ISCT), where hMSCs are defined by the positive expression of CD73, CD90, and CD105 and test negative for CD45, CD11b or CD14, CD19 or CD79a, CD34, and HLA-DR markers (5,6). The relative expression of these markers has been shown to be dependent on donor health status, the source of isolation, and laboratory processing conditions (7,8) and this, in turn, contributes error to bioprocessing quality control (QC), in addition to complicating regulatory burden. To evaluate the purity and homogeneity of isolated hMSCs, extensive efforts Cytometry Part A  00A: 0000, 2017

Original Article have been placed on flow cytometric immunophenotyping (9,10), providing a robust single cell analytical capability. However, any recorded biological variation within cell samples will be compounded by instrument properties, ambient conditions, and subjective components in data analysis inherent to these types of assays. This added variability severely hampers efforts to standardize methods for monitoring hMSC quality during bioprocessing. Standardization of polychromatic flow cytometry assays is technically challenging, although several strategies have been proposed (11–14), including FCB for increased data reproducibility between cell samples, analysts, and laboratories (15). Deploying the FCB platform in flow cytometry applications enables multiplexing of samples in combination with reduced inter-sample variability (16). To date, cell barcoding has been achieved using antibody-mediated, amine reactive cytosolic dyes, streptavidin-based, multicolor bead based, and genetic fluorescent barcoding approaches with respective advantages and disadvantages (17–21). The assay has been successfully utilized for immortal cell lines, freshly isolated (primary) cell samples, and blood cells (15,22–25). The technique tags several cell samples with unique fluorescent signatures enabling multiplexed simultaneous analyses. Various amine reactive dyes, including NHS/succinimidyl ester, carboxyfluorescein N-hydroxysuccinimidyl ester (CFSE), Celltracker Red (CTR), and CellTrace Violet (CTV) dyes have been commonly used for cytosolic FCB applications (17,19). Unlike other barcoding approaches, cytosolic barcodes are applicable to almost all cell types, and exhibit high stability (15), however, some compounds are variable with regard to signal homogeneity (i.e., CFSE cell tracker dye) (17,26) and in some cases, efficient labeling requires cell membrane permeabilization (27). So far, FCB in flow cytometry has been effectively implemented for discrimination of intracellular protein targets and has facilitated a number of applications such are highthroughput drug screening (23), cancer cell signaling profiling (24,28) as well as improved signal amplification for enhanced phosphoprotein detection (29), but assessments of surface marker expression using this assay have been limited (17,19,25,27). The current study represents a proof of principle where a method for the rapid characterization of bone marrow (BM) derived hMSCs using a modified FCB flow cytometric assay is presented together with associated technical considerations to ensure robust processing. The rationale for this method development originated from the need to monitor a number of cell surface biomarkers from multiple sources of hMSCs in a robust and economical manner. The phenotypic drift of hMSCs was monitored in serial-passaged cells expanded by multiple analysts from a common source. Three panels of grouped fluorochrome-conjugated antibodies were used to target 15 ubiquitously expressed or stage-specific surface markers in conjunction with an amine reactive, fixable viability dye FVD eFluor 506 as a barcoding dye. Assay robustness was tested by comparing data generated using nonbarcoded samples examined separately. Side-by-side analysis of barcoded and non-barcoded cells showed the potential to 2

improve reproducibility for a subset of markers when using this modified technique, which additionally can limit subjectivity during analysis of cell phenotype between analysts, labs, and under various process conditions.

MATERIALS AND METHODS Cell Culture Conditions, Enzymatic Detachment, and Cryopreservation Human bone-marrow-derived mesenchymal stromal cells (MSC-001, Rooster Bio, USA) were seeded at 4000 cells per cm2 in T175 flasks and cultured at 378C in a humidified cell culture incubator at 5% CO2, in high glucose DMEM (Gibco, UK) supplemented with 10% fetal bovine serum (FBS) (Gibco, UK) for up to five days per passage. Working cell banks of the same passage level (P1) were cryopreserved in growth medium with 10% DMSO (Sigma, UK) in liquid nitrogen (LN2) for subsequent experiments. HeLa (CCL2, ATCC, LGC Standards, UK), NT2/D1 (kindly provided by University of Sheffield, UK), HepG2 (HB-8065TM, ATCC, LGC Standards, UK), U937 (CRL-1593.2, ATCC, LGC Standards, UK), Jurkat (TIB-152, ATCC, LGC Standards, UK), and A549 (CCL-185, ATCC, LGC Standards, UK) cells were cultured according to the manufacturer’s instructions. All selected cell lines were verified to be free from mycoplasma contamination using the Mycoalert kit (Lonza, UK). For cell disassociation, two different approaches were performed using the most commonly utilized cell detachment buffers; 0.25% trypsin/EDTA (Sigma, UK) and Accutase (eBioscience, UK). Cultured BM-hMSCs were washed with DPBS and detached using selected enzymatic treatments according to the manufacturer’s instructions. Cells were cryopreserved using two commercially available cryopreservants HyClone and Synth-a-freeze (Thermo Fisher Scientific, UK) following the manufacturer’s protocol, as well as 10% DMSO (Sigma, UK) in complete growth medium. Approximately a million cells were cryopreserved in LN2 (–1908C) for further analysis at various passages (P1 > P9). Barcoding Dye Selection, Stability, Labeling Intensity, and Working Concentrations Three independent approaches were tested for cell barcoding experiments. CFSE (Thermo Fisher Scientific, UK) was used to label live cells directly by adding at 0.5 mM concentration to the reaction buffer for 20 min at 378C. Fixed BM-hMSCs were labeled using either 0.1 mg/ml (dim cells) or 1 mg/ml (bright cells) 40 ,6-diamidino-2-phenylindole (DAPI) (Sigma, UK) or alternatively fixable viability dye (FVD) eFluor 506 (eBioscience, UK) as detailed in the manufacturer’s protocol. The localized stability of FVD signal was monitored by incubating labeled cells with unstained cells for 1 h, 24 h, 7 days or one-month duration at 48C protected from light, before assessing label transfer. Dye labeling intensity was investigated under fluorescent microscopy (Nikon Eclipse TE2000, Tokyo, Japan) and flow cytometry (CantoII, BD Biosciences, UK). Immunophenotyping by fluorescent cell barcoding

Original Article Table 1. Table of grouped pre-selected surface markers, including ISCT recommended markers in conjunction with other generic and QC markers R -488/FITC ALEXA FLUORV

G1 G2 G3

CD9 (Generic) HLA-DR (ISCT –) CD24 (Generic)

PE

R -647 ALEXA FLUORV

PACIFIC BLUE/BV-421

PERCP-CY5.5.

CD166 (ISCT 1) CD29 (Generic) CD11b (ISCT –)

CD19 (ISCT –) CD45 (ISCT –) CD271 (ISCT –)

CD34 (ISCT –) CD146 (ISCT 1) CD95 (QC)

CD105 (ISCT 1) CD90 (ISCT 1) CD73 (ISCT 1)

*For an additional set of markers please see Supporting Information Table 1.

Selection of Fixation and Staining Buffers 103 BD cell fix (BD Biosciences, UK) was diluted in distilled water to make a final concentration of 13, 23, and 43 and selected surface marker expression was evaluated side by side with 4% Paraformaldehyde (PFA) fixation buffer (BioLegend, UK) as well as live/unfixed cells. In addition, the effect of 0.36% methanol (Acros organics, UK), 1% Formaldehyde (FA) methanol-free (Thermo Fisher Scientific, UK), and 1% PFA (BioLegend, UK) on antibody binding were also tested. Cells were fixed for 15 min at RT on a rotating platform, washed (33) in D-PBS (Sigma, UK) by subjecting to centrifugation (300 g for 5 min) and fixed cells were prepared for further analyses accordingly. Cell staining buffer from BioLegend (BioLegend, UK) was used for all antibody labeling procedures (unless it was specified otherwise by the manufacturer) whilst D-PBS was selected for cell barcoding. Sample Preparation and FCB Assay For FCB, cell samples were prepared as follows: cells were detached using Accutase cell detachment solution (eBioscience, UK) for 5 min at 378C in 5% CO2. Resultant cell suspensions were subjected to centrifugation at 300 g for 5 min and the residual pellet washed once in D-PBS followed by centrifugation. To assess cell viability cell pellets were resuspended in ice-cold D-PBS and labeled with the viability dye FVD eFluor 660 or FVD eFluor 780 (eBioscience, UK) following the manufacturer’s instructions. After labeling, cells were washed three times in ice-cold D-PBS by centrifuging at 300 g for 5 min at 48C. Cell pellets were resuspended in the fixation buffer (13 BD cell fix, BD Biosciences, UK) for 15 min on a rotating platform, protected from light. Fixed cells were again washed with D-PBS twice with final centrifugation at 300 g. To barcode, cells were labeled within uniform suspensions of 1 3 106 cells/ml with final FVD eFluor 506 concentrations of 0.4 ml/ml (dim) and 7 ml/ml (bright) for 15 min. In a separate approach equivalent cell suspensions were barcoded using 0.1 mg/ml (for dim cells) and 1 mg/ml DAPI (for bright cells) for 10 min (Sigma, UK) protected from light. A rotating platform was used to maintain a single cell suspension. Barcoding was also performed with CFSE dye (Thermo Fisher Scientific, UK) by labeling live cells with 0.5 mM CFSE for 20 min at 378C as per the manufacturer’s protocol (Thermo Fisher Scientific, UK). Fixed barcoded cells were washed three times in cold DPBS followed by centrifugation (300 g for 10 min) to remove the residual non-reacted dye. Following the last washing step cell pellets were resuspended in 1 ml staining buffer (BioLegend, UK). In order to achieve an equal cell number in a mix, barcoded (dim and bright) and unstained cells were Cytometry Part A  00A: 0000, 2017

enumerated using a Vi-Cell automated cell counter (Beckman Coulter ViCell XR, UK) and 50 lL cell suspensions, containing 300,000 cells per sample, were combined in a single tube. Prior to antibody labeling, barcoded cells were incubated with Human Trustain Fcx (Fc receptor blocking) buffer according to the manufacturer’s instructions (BioLegend, UK). The expression level of different surface antigens was analyzed by a direct immunolabeling approach. Cell labeling was carried out using selected groupings of pre-titrated antibody (Table 1) using the manufacturer’s instructions unless otherwise stated. Cells were incubated with antibodies for 20 min, protected from the light on a rotating platform, washed once and resuspended in 300 ml fresh D-PBS before transfer to FACS polystyrene flow cytometry tubes (BD Biosciences Falcon, UK). Samples were analyzed immediately and covered with foil during measurement sessions. A list of all reagents used during assay development is tabulated in Supporting Information Appendix 1.

FCB for analyzing multiple cell types. Barcoding of various cell types (listed in section “Cell Culture Conditions, Enzymatic Detachment, and Cryopreservation”) was performed on approximately 1 3 106 cells. All cells were fixed as in section “Sample Preparation and FCB Assay”. Two different FVDs (FVD eFluor 506 and FVD eFluor 660) were used to barcode all selected cells. Fourfold dye dilutions for each FVD were generated at 1 ll, 0.25 ll, 0.0625 ll, 0.015 ll, and 0.004 ll per ml, as previously described (15). One sample was left unstained to monitor dye transfer in the mixture. Cell barcoding time, buffers, and washing steps were carried out as in section “Sample Preparation and FCB Assay”. Immunophenotyping of Live Cells Cells were detached and prepared for immunophenotyping as in section “Cell Culture Conditions, Enzymatic Detachment, and Cryopreservation”. Live/unfixed cells following Fc blocking were labeled with selected groups of surface antibody cocktails (according to Table 1) on ice for 20 min protected from light. A single cell suspension was maintained during incubation using a rotating platform. Following incubation, the sample was washed by adding staining buffer (BioLegend, UK) and centrifuging at 300 g for 5 min at 48C. The remaining pellet was resuspending in 500 ml fresh cold D-PBS before transfer to 5 ml FACS tubes (BD Biosciences, UK) and analyzed as described below. 3

Original Article Flow Cytometry All analytical procedures were performed with either a BD FACSCanto II instrument (4–2-2 configuration optical filters) fitted with an octagon-488 nm blue laser, trigon-633 nm red laser, and trigon-405 nm violet laser, operated using FACSDiva software version 6.1.3 (BD Biosciences, UK) or a Beckman Coulter Cytoflex instrument (5–5-3 configuration optical filters), fitted with 488 nm, 638 nm, and 405 nm lasers, operated using CytExpert acquisition software (Beckman Coulter, UK). Subsequent data were saved in the list mode format and exported for analysis using FlowJo software v10.1 (Tree Star, OR). Quality control was performed daily using CS&T beads (BD Biosciences, UK). Machine sensitivity for day-to-day variation was tested using six peak SPHEROTM Ultra Rainbow fluorescent particles (Spherotech Lake Forest, IL). Compensation was set in FACSDiva software using unstained cells and cells labeled with each single color antibody as well as the highest concentration of FVD eFluor 506 dye as required. The compensation strategy was evaluated using unstained cells, cells as single color controls and cells labeled with the highest concentration of FVD eFluor 506 (Supporting Information Table 1). Automated compensation was applied in FACSDiva 6.1 acquisition software and verified during data analysis in FlowJo software to ensure that all spectral overlap was calculated correctly. The centers of scatterplots from positive and negative cell populations were aligned by matching MFIs. PMT voltages were set on unstained cells using baseline reported values as a guideline. Background staining for antibodies was determined in negative/unstained cells and with matched fluorochrome-conjugated isotype controls. Data were derived from a minimum of 10,000–20,000 events collected for each condition. Technical validation steps for both compensation and appropriate data analyses were considered throughout assay development.

Gating strategy. The following gating strategies were applied for the final data analysis: The stability of runs was checked by time versus scatter plot and debris were excluded by gating using FSC versus SSC parameters. Dead cells were further excluded by gating them out on the basis of the viability dye FVD eFluor 660 (or FVD eFluor 780) versus FSC parameter and doublets were discriminated using FSC-H versus FSC-A parameters. Fluorescence Minus One (FMO) controls were used for the gating controls and finally, barcoding deconvolution was performed using either density or contour plots, displaying a barcoding parameter FVD eFluor 506 vs. a selected marker or FSC. Gates around the three populations differentiated by FVD eFluor 506 intensity were drawn as described by Krutzik et al. (15) (Fig. 8). Immunocytochemistry Four thousand cells per cm2 of BM-MSCs were seeded in 6 well plates (Nunc, UK) at early passage level and were cultured at 378C in 5% CO2 in growth medium for 24 h. Cells were then fixed with 4% PFA (BioLegend, UK) for 10 min at RT. Fixed cells were washed three times in D-PBS and labeled corresponding to dim and bright concentrations of FVD 4

eFluor 506 for 2 h. Excess dye was removed by washing and cells were visualized under UV light with an epifluorescent microscope (Nikon Eclipse TE2000, Japan). Data and Statistical Analysis Marker expression was reported as median fluorescent intensity (MFI) from three independent samples. Statistical differences were assessed using two-way Analysis of Variance (ANOVA) with Tukey HSD post-hoc analysis. MFIs were log transformed in order to obtain the best fitted models with normally distributed residuals. All reported values represent mean 6SEM of independent MFI values.

RESULTS To analyze culture-induced changes in hMSC phenotype a single batch of primary BM-hMSCs was amplified through a total of nine passages by multiple analysts (Fig. 1) and biomarker expression was assessed at stages of early (P3), middle (P6), and late (P9) passage. Cells were characterized on the basis of ISCT recommended markers for hMSCs in conjunction with additional generic markers and those associated with QC (Table 1). It is noteworthy that no gross morphological changes were observed by bright-field microscopic examination during this period (Supporting Information Fig. 1). Overviews of the assay, as well as a breakdown of assay process and validation steps, are presented schematically (Fig. 1). An initial attempt to barcode live cells was made according to a previously described technique (17,19) using CFSE dye (0.5 mM), to assess its suitability for rapid immunophenotyping, but dye transfer to the unstained cell population was noted during subsequent antibody incubation. Although we could still discern separate cell populations following antibody labeling, it is unlikely that this dye would be suitable for the generation of more complex barcoding matrices due to the extent of dye transfer (Supporting Information Fig. 2). In a separate strategy we labeled fixed cells with different concentrations of DAPI (0.1 mg/ml and 1 mg/ml) and found that this dye was prone to even more rapid transfer than CFSE dye from the population of stained cells to unstained cells which were evident during the antibody labeling, as noted by the loss of the unstained compartment during data acquisition on a flow cytometer (Supporting Information Fig. 3). Fixable Viability Dye (FVD) eFluor 506 was next considered as an alternative barcoding candidate and initially fixed cells were labeled according to the manufacturer’s recommended concentration for viability evaluation (1 ml per 1 3 106 cells in suspension). The stability of dye localization was verified by incubating FVD stained cells with a sample of unstained cells and assessing dye transfer to this population at 1 h, 24 h, 7 days, and 1 month (Fig. 2). A clear separation on the dot-plots indicated that minimal dye transfer was evident at these time points. However, a subsequent decrease of calculated stain index (SI) values (28%, 249%, 267%, and 287.6%, respectively) was observed over this time course, when incubated at 48C protected from the light. SI was calculated as described by others (30,31). The decrease in SI value could be indicative of loss of dye compartmentalization or Immunophenotyping by fluorescent cell barcoding

Original Article

Figure 1. (A) Schematics of the FCB assay: BM-hMSC cells manufactured by three analysts (A1, A2, A3) following the same culture expansion protocol and expanded at different passage level (P1 > P9). Cells were barcoded with varying concentrations of fixable viability dye FVD eFluor 506, combined together at equal cell number with unstained cells and labeled with the corresponding antibodies. Labeled samples were processed by FACSCantoII as a single sample and analyzed accordingly. (B) Cause and effect (Ishikawa) diagram for optimizing FCB assay performance indicating technical considerations during method development.

Figure 2. Representative dot plots for FVD eFluor 506 dye stability testing. Unstained and bright stained BM-hMSC samples were mixed in a single tube at equal numbers and analyzed by flow cytometry at different time points. (A) Dot-plots represent barcoded and unstained cells immediately after mixing (red) and following 1 h (dark blue), 24 h later (purple), 7 days (cyan) and 1 months incubation (green). (B) An observed reduction in MFI and calculated SI was monitored for FVD positive samples in the mixture over the time period, but clear sample separation was still apparent following one-month incubation.

Cytometry Part A  00A: 0000, 2017

5

Original Article

Figure 3. Representative flow cytometry dot plots and epi-fluorescent microscope images to demonstrate FVD eFluor 506 dye staining intensity in fixed BM-hMSCs. Fixed cells were labeled without prior permeabilization and 7 ll FVD up-take intensity was tested by flow cytometry (using a backgating strategy). The rectangular gate represents the unstained cell (grey dot plots) boundary (panel A). FVD positive cells were gated and verified on selected scatter plots. Diffuse FVD staining (panel B) at both high (left image) and low (right image) concentration were confirmed by illumination under UV light using a 203 objective/2003 magnification (Nikon Eclipse TE2000, Japan).

changes in dye composition. Therefore, according to our data, we would recommend an analysis of barcoded samples within 24 h. To verify complete labeling of the sample without the need for cell permeabilization at discrete dye concentrations, barcoded cells were carefully investigated using both fluorescent microscopy and flow cytometry approaches. For microscopy, ten fields of view were examined at each concentration using a 203 objective (2003 magnification) and UV laser illumination. Subsequently, barcoded cells analyzed on a flow cytometer using a back-gating strategy verified that all live cell gated populations were FVD eFluor 506 positive. Both techniques confirmed that no further permeabilization

steps were required for efficient labeling using FVD eFluor 506 dye (Fig. 3). In order to barcode cell preparations from three analysts and achieve the best separation during sample deconvolution on two-dimensional plots, FVD dye concentration was titrated. A range of concentrations was examined from 0.4 to 7 ml per 1 3 106 cells (Fig. 4A). It was found that an optimum separation of the three samples was achieved using an unstained cell component (FVD eFluor 506 negative) against FVD labeled populations at 0.4 ml and 7 ml per 1 3 106 cells (Fig. 4B). To evaluate the compatibility of this technique for multiple cell types and deployment of different FVD formats

Figure 4. An example of FVD eFluor 506 dyes titration. BM-hMSC cells were fixed with 13 BD cell fix, washed and then stained with selected dilutions of the FVD dye (0.4, 0.5, 1, 2, 3, 4, 5, and 7 ml). After staining was complete, cells were washed three times by subjecting to centrifugation and analyzed by flow cytometry. Panel A shows histograms at indicated concentrations of FVD eFluor 506 dye and Panel B demonstrates an optimal concentration of FVD eFluor 506 to achieve effective separation between FVD negative and positive samples. (MFI fold changes are presented in Supporting Information Table 2).

6

Immunophenotyping by fluorescent cell barcoding

Original Article

Figure 5. Barcoded U937 cells were distinguished by using modified FCB method with FVD as a barcoding marker. U937 cells were labeled with five concentrations of FVD eFluor 660 and FVD eFluor 506 (1 ll, 0.25 ll, 0.0625 ll, 0.016 ll, and 0.004 ll.) and combined with unstained cells. Plotting the FCB channels (FVD eFluor 506 and FVD eFluor 660) versus forward scatter (FSC) (A) or each other (B), reveals six distinct populations that correspond to the individual barcoded sample tube. Once gated, each sample can be analyzed accordingly. (MFI fold changes are presented in Supporting Information Table 2).

(FVD eFluor 506 and FVD eFluor 660) with the intention of generating more complex barcoding matrices, various cell types including Jurkat (0.06 ll FVD eFluor 660), Hela (0.06 ll FVD eFluor 506) A549 (0.25 ll FVD eFluor 506/660), NT2/ D1 (1 ll FVD eFluor 506/660) and hMSC (4 ll FVD eFluor 660) were labeled and combined with unstained Jurkat cells (Supporting Information Fig. 4 A). Further evaluation was sort using a different combination of cells and adjusted dye concentrations; HeLa (0.125 ll FVD 506 and 0.03 ll FVD 660), A549 (0.25 ll FVD 506 and 0.125 ll FVD 660), U937 (1 ll FVD 506 and 0.5 ll FVD 660) cells and unstained HepG2 were analyzed in a mixed sample (Supporting Information Fig. 4B). In addition, we tested the suitability and efficiency of FVD dyes as a barcoding agent on U937 cells, which were originally used for the barcoding method development by Krutzik and Nolan (15). U937 cells were barcoded using a fourfold dilution of each selected FVD dye separately and 6 populations were generated per FVD marker, including unstained. In two separate barcoding approaches utilizing different cell types or U937 cells, each sample was easily identifiable on scatter plots after deconvolution (Fig. 5, Supporting Information Fig. 4). However, the optimization of dye concentration for individual cell types should be assay specific. It was observed that thorough washing steps following labeling were also essential to minimize the transfer of residual/

unwashed dye from the high concentration labeled compartment (5 ll/ml) to non-labeled cells. Omitting just one washing step accounted for residual dye transfer to the negative cell compartment (Supporting Information Fig. 7). Utilization of FVD eFluor 506 as a barcoding marker requires cell fixation prior to immunophenotyping. Fixation conditions have a considerable effect on the performance of labeled antibodies for flow cytometry as well as the expression of target antigens (32–34) and should be carefully considered during assay development. The reproducibility of fixation conditions was assessed within method development using a range of commercially obtained fixatives (PFA/FA), measuring marker expression side-by-side with intact/unfixed cells. Each cell preparation was subjected to staining for MSC -positive (ISCT), generic and QC markers and subsequently analyzed by flow cytometry (Supporting Information Table 1). No single fixation condition was shown to support optimal expresR 488sion of all selected markers (PE-CD73, Alexa FluorV R CD105, Pacific blue-CD95, APC/Cy7-CD29, Alexa FluorV 647-CD44, and Pecp-Cy5.5-CD90) (Table 2), but on the basis of these results a 13 concentration of BD CellFIX buffer (BD Biosciences) gave rise to expression levels that most resembled live cells for the majority of ISCT positive markers. It is likely that further optimization may be achieved for certain targets by exploring alternative fixative methods in side-by-side

Table 2. Resultant expression of ISCT recommended hMSC, generic and QC markers under indicated fixation conditions as a percentage of live-cell (unfixed) recorded level. Percentages are given as a mean with the respective standard deviation (SD) indicated in parentheses (n 5 3). MARKERS

MARKER TYPE

CD105 CD44 CD29 CD95 CD73 CD90

ISCT 1 Generic Generic QC ISCT 1 ISCT 1

13 BD CELL FIX (BD BIOSCIENCE)

23 BD CELL FIX

43 BD CELL FIX

4% PFA (BIOLEGEND)

LIVE (CONTROL)

49.93(2.78) 73.89(0.99) 80.68(1.95) 88.92(0.96) 34.33(1.04) 47.19(2.35)

68.75(3.69) 72.36(0.98) 63.26(3.47) 101.71(4.7) 17.76(0.52) 38.55(3.48)

27.93(2.92) 85.58(6.26) 83.34(4.57) 90.44(4.12) 6.13(0.28) 55.99(7.22)

35.18(3.09) 78.27(1.33) 83.84(4.34) 88.56(7.30) 14.77(0.63) 50.04(3.10)

100 100 100 100 100 100

Cytometry Part A  00A: 0000, 2017

7

Original Article

Figure 6. Fixation did not dramatically alter marker expression patterns. The X–Y plot of two different approaches was used to evaluate the surface antigen expression on BM-hMSCs. The comparison of live and fixed cell (modified FCB) immunophenotyping methods indicate positive correlation (R2 5 0.9346). Data presented as mean of MFI (Log 10) values of positive cells (n 5 3). [Color figure can be viewed at wileyonlinelibrary.com]

comparative studies (Supporting Information Fig. 5). Antibody (Ab) binding efficacy after fixation was assessed by measuring the percentage of positive cells as described in Ref. 41. The selected fixative was then tested within an analysis of all designated markers (Table 1) expressed on intact/unfixed cells at the early passage level. Despite the fact that mild fixation was found to alter the expression of some generic (CD29, CD24, CD95), ISCT positive (CD90, CD105, CD166), and ISCT negative (CD34) markers, it was found that the general expression trend was not modified (Supporting Information Fig. 6). In addition, the suitability of this modified FCB assay for immunophenotyping was confirmed by noting a positive correlation against live cell surface marker expression (Fig. 6). After selecting the most appropriate assay conditions and in order to validate the optimized protocol, the expression of 15 pre-selected surface markers were investigated and analyzed in co-staining groups (Table 1) throughout the BM-hMSC culture expansion process. Barcoding was performed using approximately 1 3 106 cells prepared by independent analysts with cell preparations labeled A1, A2, A3 randomly assigned for barcoding at either low or high concentration of eFluor 506 against the remaining unstained sample. An equal number of A1-A3 barcoded cells (approximately 300,000) were combined in suspension and analyzed as a single sample (Fig. 1). The effect of different target, passage and analyst was tested by ANOVA and no significant differences between the three analysts were found. On removing analyst (as a variable) and reanalyzing the data, very strongly significant effects (P  0.0001) were found for both passage and surface markers, as well as a strong interaction between the two (Supporting Information Fig. 10). This result was obtained for all three co8

staining groups (ISCT negative, ISCT positive, and Generic/ QC). Interestingly, some ISCT positive and generic markers were found to be markedly upregulated in cells at the middle passage (P6) compared to early (P3) and late (P9) passage including, CD29, CD90, and CD166 (Fig. 7). The developed assay was deployed to explore any resultant impact on measured BM-hMSC phenotype by modification of critical bioprocessing conditions, including hMSC harvesting techniques and cell cryopreservation methodologies (35). Harsh enzymatic cell detachment has a strong influence on surface marker expression as well as on cell viability (36) and this process should be conducted with appropriate buffering. It is important that hMSCs harvested using these methods remain viable with identity biomarkers expressed within a specified range. The most broadly used cell dissociation reagents (0.25% trypsin/EDTA and Accutase) were compared side-by-side using the FCB approach (Supporting Information Fig. 8) to evaluate cell biomarker expression patterns. In a similar manner, changes in the marker expression level were also investigated in hMSCs cryopreserved using two commercially available cryopreservants (HyClone and Syntha-freeze) as well as an in-house preparation of 10% DMSO in complete media. It was found that no changes to selected R 488-CD105, Pacific bluemarker (PE-CD73, Alexa FluorV R 647-CD44, and CD95/CD34, APC/Cy7-CD29, Alexa FluorV Pecp-Cy5.5-CD90) expression were apparent using any of these conditions (Supporting Information Fig. 9).

DISCUSSION Currently, flow cytometry is the method of choice for stem cell immunophenotyping (37–39), where it has great potential Immunophenotyping by fluorescent cell barcoding

Original Article

Figure 7. Certain biomarkers are sensitive to modulation over culture expansion (early, middle and late passages) but comparable levels are observed between analysts at the same passages. Employing the FCB assay demonstrates the improved reproducibility. Biomarker expression level assessed by flow cytometric median fluorescent intensity (MFI) values (n 5 3). (Error bars denote 6SEM). (Data are presented on log scale to amplify negative marker MFI values for better visibility). [Color figure can be viewed at wileyonlinelibrary.com]

for bioprocessing due to its single-cell analysis format and inline compatibility. At present, efforts to identify combinations of cell surface biomarkers that may provide more accurate QC metrics for cell manufacturing, particularly applicable to heterogeneous cell populations, such as isolated hMSCs are ongoing. While it is broadly accepted that a large proportion of observed error in sample data is attributed to biological variation, the contribution of technical error can be considerable and for flow cytometry this will comprise subjectivity in data analysis, correct instrument settings and sample processing or diurnal changes in instrument sensitivity. Minimizing this technical variation is a challenge for method standardization but can ensure accuracy in reported data. Utilizing an FCB assay in flow cytometry diminishes some of these potential processing errors, however, earlier reported FCBs were focused on multiplex analyses of intracellular signaling mechanisms (e.g., studies of protein phosphorylation). Unfortunately, these assays require alcohol permeabilization steps (15) making them unsuitable for surface marker characterization. In 2014, Behbehani and others (27) presented a modified assay utilizing transient partial 0.02% saponin permeabilization prior to barcoding, making it more applicable to robust surface marker analysis. The choice of compatible dye chemistry for cytosolic FCB is also critical for success. Previously reported studies have relied on N-hydroxysuccinimide (NHS) ester labeling primary amine (R-NH2) groups or other aminecontaining molecules, making this dye a good barcoding candidate and suitable for all cell types, however the extent of cell permeabilization was found to be a critical parameter in some studies for surface marker labeling, leading to challenging optimization requirements (15,16,27). Further examples of modified cytosolic FCB assays using a selective combination Cytometry Part A  00A: 0000, 2017

of amine reactive cell tracking dyes (CFSE and CTV or CTR) in cancer marker studies have been reported, and common issues for achieving reproducible labeling using the abovementioned dyes were considered (26). On the other hand, alternative barcoding techniques such as antibody- or bead conjugates have not been reported to be compatible with all cell types and establishing routine generic barcoding techniques in the laboratory is time-consuming and costly (18–20). In the current article, we aimed to design a modified fluorescent barcoding assay for rapid surface marker analyses eliminating the need for a specific cell permeabilization step, as well as potentially improving reproducibility. This was accomplished utilizing the fixable viability dye (FVD) eFluor 506 for the first time in a barcoding technique. The rationale for selecting FVD eFluor 506 as a potential barcoding agent was simply based on its physical/optical properties (emission spectrum 506 nm). This particular format can be detected by a 510/50 band pass filter (the equivalent of Amcyan) and unlike other cytosolic barcoding, dyes do not engage considerably “busier” channels. This may be beneficial during FCB multicolor panel design for instruments employing a similar optical configuration as the BD Canto II (4–2-2) instrument up to eight color limitations. This viability dye belongs to the amine reactive dye Group (40) and transits the compromised membranes of dead cells, reacting with free amines intracellularly as well as on the cell surface, yielding a high fluorescence signal. As a consequence, the incorporation of a mild fixation step prior to cell barcoding renders all cells accessible to the dye, meaning that a fluorescent signature (dim or bright) is based solely on the selected dye concentration used in the assay. As a proof of concept in this article, a simple barcoding matrix was generated, but the assay has the potential for more complex multi-parametric applications 9

Original Article

Figure 8. Representative gating strategy: debris (A), dead cells (B), and doublets (C) were gated out. Stability of runs was verified by time versus parameter (D). FMO gating are presented on (E) and (F) plots demonstrating examples of CD9 and CD90 gating and FMO boundary (gray dot plots/rectangular gate E-minus FITC conjugated antibody, F-minus PerCp-Cy5.5 conjugated antibody from selected groups G1 and G2-Table 1). The FMO helps to resolve the ambiguity of a positive signal. On AmCyan (barcoding) channel plotted against FSC parameter three barcoded samples were evident. Deconvolution helps to identify different analyst samples in a mixture (G), can be easily gated and the selected marker expression in each sample analyzed simultaneously by measuring the median fluorescent intensities (H). [Color figure can be viewed at wileyonlinelibrary.com]

utilizing mixtures of different formats of FVD eFluor dyes or by up- or down-titration of particular FVD formats (Fig. 5, Supporting Information Fig. 4). An assessment of the suitability of this modified FCB approach for intracellular protein analyses is ongoing. Cell fixation prior to barcoding may convey associated advantages or disadvantages. Although it minimizes rapid internalization of surface markers during lengthy immunophenotypic assays it may also alter antibody binding epitopes or disrupt membrane integrity with resultant measurement error (32–34). To estimate the detrimental effect of nonoptimized fixation conditions during the FCB procedure, a comprehensive phenotypic analysis of live cells was used as a 10

starting point for assay development. These live-cell marker expression levels were used as a guideline for the selection of appropriate assay conditions including the optimal concentration of PFA for fixation. Overall, this approach permitted the rapid phenotypic assessment of cultured hMSCs exposed to various bioprocessing conditions during culture expansion over nine passages by three independent analysts. The strategy of mixing barcoded cell samples for common analysis by flow cytometry, diminished some subjective factors associated with the flow cytometry application, such as adjustment of correct voltages, sample preparation and gating strategies. In addition, the assay eliminated the effect of dayto-day variation triggered by the instrument (e.g., changes in Immunophenotyping by fluorescent cell barcoding

Original Article Qr and Br values) and the need for extensive standardization. Therefore, FCB offers some advantages conferring data quality and accuracy and has the potential to reduce data misinterpretation between independent experiments. As a consequence, this modified FCB assay may be more suited to detecting subtle alterations in marker expression level during cell product quality assessments compared to conventional flow cytometry approaches. Specifically, these combined benefits result in an improved immunophenotyping protocol for QC monitoring of hMSCs which may be transferable to a variety of cell bioprocessing applications.

ACKNOWLEDGMENTS The authors would like to thank Dr Gary Morley and Dr Julie T. Davies for providing cell culture assistance and Dr Simon Cowen for assistance with statistical analysis. This work received a first place poster prize at the FlowCytometryUK 2016 meeting in Leeds.

LITERATURE CITED 1. Gebler A, Zabel O, Seliger B. The immunomodulatory capacity of mesenchymal stem cells. Trends Mol Med 2012;18:128–134. 2. Sharma RR, Pollock K, Hubel A, McKenna D. Mesenchymal stem or stromal cells: A review of clinical applications and manufacturing practices. Transfusion 2014;54: 1418–1437. 3. Petrie Aronin CE, Tuan RS. Therapeutic potential of the immunomodulatory activities of adult mesenchymal stem cells. Birth Defects Res C Embryo Today 2010;90:67–74. 4. Wei X, Yang X, Han Z-P, Qu F-F, Shao L, Shi Y-F. Mesenchymal stem cells: A new trend for cell therapy. Acta Pharmacologica Sinica 2013;34:747–754. 5. Dominici M, Le Blanc K, Mueller I, Slaper-Cortenbach I, Marini F, Krause D, Deans R, Keating A, Prockop D, Horwitz E. Minimal criteria for defining multipotent mesenchymal stromal cells. The International Society for Cellular Therapy position statement. Cytotherapy 2006;8:315–317. 6. Pilz GA, Braun J, Ulrich C, Felka T, Warstat K, Ruh M, Schewe B, Abele H, Larbi A, Aicher WK. Human mesenchymal stromal cells express CD14 cross-reactive epitopes. Cytometry Part A 2011;79A:635–645. 7. Williams DJ, Archer R, Archibald P, Bantounas I, Baptista R, Barker R, Barry J, Bietrix F, Blair N, Braybrook J, et al. Comparability: Manufacturing, characterization and controls, report of a UK Regenerative Medicine Platform Pluripotent Stem Cell Platform Workshop, Trinity Hall, Cambridge, 14?15 September 2015. Regen Med 2016;11:483–492. 8. Siegel G, Kluba T, Hermanutz-Klein U, Bieback K, Northoff H, Sch€afer R. Phenotype, donor age and gender affect function of human bone marrow-derived mesenchymal stromal cells. BMC Med 2013;11:146. 9. Paebst F, Piehler D, Brehm W, Heller S, Schroeck C, Tarnok A, Burk J. Comparative immunophenotyping of equine multipotent mesenchymal stromal cells: An approach toward a standardized definition. Cytometry Part A 2014;85A:678–687. 10. Tarnok A, Ulrich H, Bocsi J. Phenotypes of stem cells from diverse origin. Cytometry Part A 2010;77A:6–10. 11. Maecker HT, Trotter J. Flow cytometry controls, instrument setup, and the determination of positivity. Cytometry Part A 2006;69A:1037–1042. 12. Baradez M-O, Lekishvili T, Marshall D. Rapid phenotypic fingerprinting of cell products by robust measurement of ubiquitous surface markers. Cytometry Part A 2015; 87A:624–635. 13. Kalina T, Flores-Montero J, van der Velden VHJ, Martin-Ayuso M, B€ ottcher S, Ritgen M, Almeida J, Lhermitte L, Asnafi V, Mendonc¸a A, EuroFlow Consortium (EU-FP6, LSHB-CT-2006–018708), et al. EuroFlow standardization of flow cytometer instrument settings and immunophenotyping protocols. Leukemia 2012;26: 1986–2010. 14. Maecker HT, Rinfret A, D’Souza P, Darden J, Roig E, Landry C, Hayes P, Birungi J, Anzala O, Garcia M, et al. Standardization of cytokine flow cytometry assays. BMC Immunol 2005;6:13. 15. Krutzik PO, Nolan GP. Fluorescent cell barcoding in flow cytometry allows highthroughput drug screening and signaling profiling. Nat Meth 2006;3:361–368. 16. Krutzik PO, Clutter MR, Trejo A, Nolan GP. Fluorescent cell barcoding for multiplex flow cytometry. Curr Protoc Cytom 2011;Chapter 6:Unit 6.31.

Cytometry Part A  00A: 0000, 2017

17. Yamanaka YJ, Szeto GL, Gierahn TM, Forcier TL, Benedict KF, Brefo MSN, Lauffenburger DA, Irvine DJ, Love JC. Cellular barcodes for efficiently profiling single-cell secretory responses by microengraving. Anal Chem 2012;84:10531–10536. 18. Akkaya B, Miozzo P, Holstein AH, Shevach EM, Pierce SK, Akkaya M. A simple, versatile antibody-based barcoding method for flow cytometry. J Immunol 2016;197: 2027–2038. 19. Sukhdeo K, Paramban RI, Vidal JG, Elia J, Martin J, Rivera M, Carrasco DR, Jarrar A, Kalady MF, Carson CT, et al. Multiplex flow cytometry barcoding and antibody arrays identify surface antigen profiles of primary and metastatic colon cancer cell lines. PLoS One 2013;8:e53015–e53019. 20. Davies R, Vogelsang P, Jonsson R, Appel S. An optimized multiplex flow cytometry protocol for the analysis of intracellular signaling in peripheral blood mononuclear cells. J Immunol Methods 2016;436:58–63. 21. Smurthwaite CA, Hilton BJ, O’Hanlon R, Stolp ZD, Hancock BM, Abbadessa D, Stotland A, Sklar LA, Wolkowicz R. Fluorescent genetic barcoding in mammalian cells for enhanced multiplexing capabilities in flow cytometry. Cytometry Part A 2014;85A:105–113. 22. Stam J, Abdulahad W, Huitema MG, Roozendaal C, Limburg PC, van Stuijvenberg M, Sch lvinck EH. Fluorescent cell barcoding as a tool to assess the age-related development of intracellular cytokine production in small amounts of blood from infants. PLoS One 2011;6:e25690–e25698. 23. Krutzik PO, Crane JM, Clutter MR, Nolan GP. High-content single-cell drug screening with phosphospecific flow cytometry. Nat Chem Biol 2007;4:132–142. 24. Irish JM, Myklebust JH, Alizadeh AA, Houot R, Sharman JP, Czerwinski DK, Nolan GP, Levy R. B-cell signaling networks reveal a negative prognostic human lymphoma cell subset that emerges during tumor progression. Proc Natl Acad Sci USA 2010; 107:12747–12754. 25. Giudice V, Feng X, Kajigaya S, Biancotto A, Young NS. Fluorescent Cell barcoding as new flow cytometric technique for multiplexed phenotyping and signaling profiling in hematologic patients. Blood 2016;128:5033. 26. Tario JD, Jr., Humphrey K, Bantly AD, Muirhead KA, Moore JS, Wallace PK. Optimized staining and proliferation modeling methods for cell division monitoring using cell tracking dyes. J Vis Exp 2012;1–11. 27. Behbehani GK, Thom C, Zunder ER, Finck R, Gaudilliere B, Fragiadakis GK, Fantl WJ, Nolan GP. Transient partial permeabilization with saponin enables cellular barcoding prior to surface marker staining. Cytometry Part A 2014;85A:1011–1019. 28. Ullal AV, Peterson V, Agasti SS, Tuang S, Juric D, Castro CM, Weissleder R. Cancer cell profiling by barcoding allows multiplexed protein analysis in fine-needle aspirates. Sci Trans Med 2014;6:219ra9. 29. Clutter MR, Heffner GC, Krutzik PO, Sachen KL, Nolan GP. Tyramide signal amplification for analysis of kinase activity by intracellular flow cytometry. Cytometry Part A 2010;77A:1020–1031. 30. Maecker HT, Frey T, Nomura LE, Trotter J. Selecting fluorochrome conjugates for maximum sensitivity. Cytometry Part A 2004;62A:169–173. 31. Baumgarth N, Roederer M. A practical approach to multicolor flow cytometry for immunophenotyping. J Immunol Methods 2000;243:77–97. 32. McCarthy DA, Macey MG, Cahill MR, Newland AC. Effect of fixation on quantification of the expression of leucocyte function-associated surface antigens. Cytometry 1994;17:39–49. 33. Cahill MR, Macey MG, Newland AC. Fixation with formaldehyde induces expression of activation dependent platelet membrane glycoproteins, P selectin (CD62) and GP53 (CD63). Br J Haematol 1993;84:527–529. 34. Schmidt V, Hilberg T, Franke G, Gl€aser D, Gabriel HHW. Paraformaldehyde fixation induces a systematic activation of platelets. Platelets 2003;14:287–294. 35. Franc¸ois M, Copland IB, Yuan S, Romieu-Mourez R, Waller EK, Galipeau J. Cryopreserved mesenchymal stromal cells display impaired immunosuppressive properties as a result of heat-shock response and impaired interferon-c licensing. Cytotherapy 2012;14:147–152. 36. Huang H-L, Hsing H-W, Lai T-C, Chen Y-W, Lee T-R, Chan H-T, Lyu P-C, Wu C-L, Lu Y-C, Lin S-T, Lin C-W, Lai C-H, Chang H-T, Chou H-C, Chan H-L. Trypsininduced proteome alteration during cell subculture in mammalian cells. J Biomed Sci 2010;17:36. 37. Donnenberg VS, Ulrich H, Tarnok A. Cytometry in stem cell research and therapy. Cytometry Part A 2012;83A:1–4. 38. Gedye CA, Hussain A, Paterson J, Smrke A, Saini H, Sirskyj D, Pereira K, Lobo N, Stewart J, Go C, et al. Cell surface profiling using high-throughput flow cytometry: A platform for biomarker discovery and analysis of cellular heterogeneity. PLoS One 2014;9:e105602–e105614. 39. Chow S, Hedley D, Grom P, Magari R, Jacobberger JW, Shankey TV. Whole blood fixation and permeabilization protocol with red blood cell lysis for flow cytometry of intracellular phosphorylated epitopes in leukocyte subpopulations. Cytometry Part A 2005;67A:4–17. 40. Perfetto SP, Chattopadhyay PK, Lamoreaux L, Nguyen R, Ambrozak D, Koup RA, Roederer M. Amine-reactive dyes for dead cell discrimination in fixed samples. Curr Protoc Cytom 2010;Chapter 9:Unit 9.34. 41. Krutzik PO, Clutter MR, Nolan GP. Coordinate analysis of murine immune cell surface markers and intracellular phosphoproteins by flow cytometry. J Immunol 2005; 175:2357–2365.

11