SEARCH

SEARCH BY CITATION

Keywords:

  • tandem conjugates;
  • monoclonal antibody cocktails;
  • spectral spillover;
  • compensation matrix

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. LITERATURE CITED
  9. Supporting Information

Background

The stability and performance of tandem-conjugated antibodies can be impaired when stored in antisera cocktails (Biancotto et al., J Immunol Methods 2011;363:245–261; Rawstron et al., Leukemia 2013;27:142–149). This, and the need for frequent re-compensation due to the possible spectral spillover variation between tandem lots, reduces the robustness of clinical flow cytometry panels that include tandems. Since tandems are required for standard 8–10 color screens, further studies of the stability of tandems in cocktails and their spillover variability are warranted.

Methods

The performance of PE- and APC-tandems stored in cocktails was tested on fresh bone marrow, preserved blood and lyophilized cell samples over 1-, 6-, or 8-week periods, respectively, and their spillover matrices were compared. The observed correction factor differences were used as the basis for analyzing how the application of an incorrect compensation matrix could influence data interpretation.

Results

Signal intensities and background fluorescence remained constant for all fluorochromes in the cocktails tested. Spillover correction factors for different PE-Cy7 mAbs did not exceed or were only marginally higher than those for non-tandem organic dye-conjugated mAb. By applying the correction factor differences observed between tandem mAb lots to clinical data, it was found that the over and under compensation would not alter the clinical interpretation.

Conclusions

Tandems can be safely stored and used in cocktails. However, each cocktail should be tested on relevant material prior to use. Exact compensation settings are a requirement for accurate data. Provided that careful evaluation of tandem compensation requirements is carried out, certain tandems may use a generic compensation matrix. © 2013 International Clinical Cytometry Society


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. LITERATURE CITED
  9. Supporting Information

Tandem conjugates (‘tandems') for use in multicolor flow cytometry (MFC) were first introduced more than 20 years ago [1, 2]. Tandems comprise two fluorochromes: one donor and one receiver. The fluorescence resonance energy transfer efficiency between the donor and the receiver may vary between batches of a particular tandem. Hence, the spectral spillover characteristics can differ from lot-to-lot [3-5]. Therefore, it is generally recommended that each tandem lot is re-compensated [4, 6-8]. It is important to not expose tandems to light, or other conditions that can break the coupling between the two dyes, since a de-coupled donor component will emit at its usual wavelength range instead of the longer wavelengths expected for the tandem [7]. Storage in dark bottles increases the durability of tandems [3]. On the other hand, the performance of tandem mAbs that are stored in antisera cocktails has been questioned. One study showed that the fluorescence signals of two q-dot-conjugated mAbs and that of a tandem-conjugated mAb were significantly reduced after storage for 3 days in a cocktail [1].

These issues pose concerns about the use of tandem conjugates in clinical laboratories, where the antibody turn-over is high. If tandems are selected for commonly used antibodies they may be used several times daily. This could potentially accelerate light-induced degradation. The use of antibody cocktails is often favored for 6–10 color screens, as these help speed up the work and prevent the risk of adding the wrong antisera or volume. The inability to use tandems in cocktails would hamper the work-flow in the lab. In addition, due to the high antibody turnover, compensation of new lots would be required very frequently. Although this is a straightforward procedure, most diagnostics laboratories are very busy and use several tandems that each requires individual compensation. In addition, the number of appropriately cytometry-trained staff can be low. Consequently, it is highly likely that new lots are introduced without prior compensation. Subsequent off-line compensation could be applied, but this strategy may cause an unacceptable delay in obtaining diagnostic results, and may also add to the quality control work load.

For these reasons, the use of tandem conjugates would reduce the overall robustness of diagnostic MFC. Nonetheless, the use of MFC for haematology–oncology diagnostics has proven very useful and, until suitable alternative fluorochromes are available, tandem conjugates are needed to achieve MFC. In addition, the last few years have seen an increase in the use of 405 nm (‘violet') lasers in clinical laboratories. Several different violet excited organic fluorochromes are commercially available. Most have emission maxima around 420–455 nm (‘blue channel') or 500–530 nm (‘yellow channel'). Violet laser-excited tandems are commercially available. However, the majority of mAb excited by the violet laser that are in clinical use today are not tandems and therefore not expected to cause particular concerns with regard to lot-to-lot variations.

Here, we examine the use of tandem-conjugated antibodies from a hematology–oncology diagnostics laboratory perspective. We analyzed the performance of tandems stored in cocktails, their lot-to-lot spectral spillover variation, and how incorrect compensation matrices may influence data interpretation. Since violet laser exited fluorochromes are relatively new in clinical laboratories, we also looked at the spectral spillover characteristics of different organic fluorochromes exited by violet lasers.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. LITERATURE CITED
  9. Supporting Information

Stability of Tandems in Antibody Cocktails Used to Label Preserved Blood and Fresh Bone Marrow

Reagents

Cytometer Setup & Tracking (CS&T) Beads, CompBeads, and Multi-Check Control whole blood (‘Multi-Check Control') were obtained from BD Biosciences. CytoComp lyophilized cells were from Beckman Coulter. Monoclonal antibodies for cocktail 1: CD4-fluorescein isothiocyanate (FITC) (SK3), CD8-Phycoerythrin (PE) (SK1), CD19-peridinin-chlorophyll α complex-cyanine5.5 (PerCP-Cy5.5) (J25C1), CD2-PE-cyanine7 (PECy7) (S5.2), CD20-allophyocyanine (APC) (L27), CD3-APC-Cy7 (SK7), CD56-V450 (B159), and CD45-V500 (2D1), all from BD Biosciences. Kappa-FITC (Rabbit polyclonal Ab) and Lambda-PE (Rabbit polyclonal Ab) were from Alere.

Method
Preserved blood and lyophilized cells

The cocktail was stored in a dark glass bottle at 4°C. The Multi-Check Control was stored and handled according to the manufacturer's instructions, and aliquots were removed weekly for labeling. CytoComp cells were reconstituted according to the manufacturer's instructions. A new vial of cells was reconstituted each week prior to use, and the same batch of cells was used throughout the study. The antibody cocktail was used to label cells on the day of preparation and subsequently on a weekly basis for 6 (Multi-Check) or 8 (CytoComp cells) weeks. To mimic daily use and light exposure, the cocktails were placed on the bench, on 4°C cool-blocks, with the cap removed for 10 minutes each day for 5 days/week throughout the test period. The time periods were deemed to reflect the combined time that cocktail bottles are opened when used several times a day.

Fresh bone marrow

A second cocktail consisting of CD16-FITC (3G8), CD117-PC7 (95C3), HLA-DR-Pacific Blue (PB, Immu-357, all from Beckman Coulter), CD13-PE (L138), CD34-PerCp-Cy5.5 (8G12), CD33-APC (P67.6), CD11b-APC-Cy7 (ICRF44), and CD45-V500 (2D1, all from BD Biosciences) was made and stored as above. A fresh (2–5 hours old) bone marrow sample was stained with this cocktail, or the same antibodies but individually pipetted, for five consecutive days. Daily light exposure was as above but for 20 minutes each day.

Flow cytometry

A Canto II flow cytometer with standard laser and filter sets and FACSDiva software version 6.1.3 was used. The instrument was calibrated daily with CS&T beads and the analysis protocol used the daily corrected CS&T bead PMT voltages to ensure consistency of fluorescence signals over the 8-week period. Compensation was set using CompBeads and the Diva-automated compensation program according to the manufacturer's instructions.

Analysis
Preserved blood and lyophilized cells

Cells were identified on a CD45-V500 (x-axis, logarithmic scale) and side scatter (SSC) (y-axis, linear scale) plot. CD45+ cells were displayed on a forward scatter (FSC) (x-axis, linear scale) and SSC (y-axis, linear scale) plot. For the first cocktail, lymphocytes were gated and displayed on plots as follows: CD2-PE-Cy7 (x-axis) and CD3-APC-Cy7 (y-axis); and CD19-PerCP-Cy5.5 (x-axis) and CD20-APC (x-axis). A CD4-FITC (x-axis) and CD8-PE (y-axis) plot was gated on CD3+ cells. A Kappa-FITC (x-axis) and Lambda-PE (y-axis) plot was gated on CD20+ cells. CD56+ cells were identified in a plot of CD56-V450 (x-axis) and CD2-PE-Cy7 (y-axis) gated on all lymphocytes.

Fresh bone marrow

Data were analyzed using Kaluza software. Neutrophils and monocytes were gated in a SSC (y-axis) CD33 (x-axis) plot. Neutrophils were then displayed on CD13 (y-axis) CD11b (x-axis) and CD13 (y-axis) CD16 (x-axis) plots and gates set around the most mature (Step 4) and immature (Step 1 or Step 2) populations [9-11]. Monocytes were displayed in a DR (y-axis) 11b (x-axis) plot, and immature cells gated as CD11bdim/negative HLA-DRbright [9-12]. A gate around the bone marrow progenitor cell compartment [13] extended to include erythroid cells was set in a SSC-A (y-axis) CD45 (x-axis) plot. From this, CD34+CD117+ myeloid progenitors were gated in a CD117 (y-axis) CD34 (x-axis) plot and CD117+HLA-DR- erythroid and myeloid progenitors from a CD117 (y-axis) HLA-DR (x-axis) plot. All fluorochrome plots used logarithmic scales. The percentage positive and median fluorescence intensity for each fluorochrome was then recorded.

Comparison of Compensation Matrices for mAbs Conjugated to the Same Fluorochrome

Monoclonal antibodies

FMC7-FITC (FMC7), CD2-FITC (S5.2), CD38-FITC (HB7), CD57-FITC (HNK-1), CD81-FITC (JS-81),CD103-FITC (Ber-ACT8), CD8-PE (SK1), CD13-PE (L138), CD14-PE (MФP9), CD56-PE, (NCAM16.2), CD59-PE (p282H19), CD79a-PE (HM47), CD123-PE (9F5), CD4-PerCP (SK3), CD19-PerCP (4G7), CD34-PerCP (8G12),CD-5PE-Cy7 (L17F12), CD10-PE-Cy7 (HI10a), CD14-PE-Cy7 (M5E2), CD19-PE-Cy7 (SJ25C1), CD25-PE-Cy7 (2A3), CD38-PE-Cy7 (HB7), CD117-PE-Cy7 (104D2), CD5-APC (L17F12), CD8-APC (RPA-T8), CD11c-APC (S-HCL-3), CD23-APC (EBVCS-5), CD33-APC (P67.6), CD55-APC (IA10), CD138-APC (MI15), CD45-V450 (H130), CD3-APC-Cy7 (SK7), CD20-APC-Cy7 (L27), CD45-APC-Cy7 (2D1), and HLA-DR-APC-H7 (L243), were from BD Biosciences. CD11b-FITC (CDIS1/18) was from Caltag Medsystems. MPO-FITC (CLB-MPO-1) was from Alere/Dako and CD160-PE (By55) was from Beckman Coulter.

Method

Seven combinations of seven antibodies were selected, each containing antibodies conjugated to FITC, PE, PerCP, PECy7, APC, APCCy7, and V450 (Supporting Information Table 1). The combinations were then used to label CompBeads according to the manufacturer's instructions.

Table 1. Analysis of the Percentage Positive Cells and the Median Fluorescence Intensity (MFI) of Six Fluorochromes in a Cocktail Used to Label CytoComp Cells Over an 8-Week Period and Multi-Check Control Cells over a 6-Week Period
 CytoComp cellsMulti-check control cells
 % PositiveMFI% PositiveMFI
mAbMean (SD)CVMean (SD)CVMean (SD)CVMean (SD)CV
CD2-PE-Cy781.2 (1.2)1.53,110 (258)8.378.0 (1.5)2.04,836 (445.1)9.2
CD3-APC-Cy776.2 (0.8)1.09,543 (537)5.674.2 (1.6)2.27,006 (503)7.2
CD4-FITC60.5 (0.8)1.42,434 (166)6.857.1 (1.0)1.86,248 (501)8.1
CD8-PE30.9 (0.6)1.921,024 (1,408)6.730.3 (0.6)1.833,522 (2,626)7.8
CD19-PerCP-Cy5.59.5 (0.5)5.0845 (55)6.513.9 (0.6)4.22,571 (240)9.3
CD20-APC9.4 (0.4)3.815,158 (1,471)9.714.1 (0.7)4.917,309 (1,288)7.4
CD56-V450n.a n.a 1.1 (0.1)12.34,077 (968)23.7
Kappa-FITCn.a n.a 55.6 (1.7)3.012,723 (561)4.4
Lambda-PEn.a n.a 39.1 (21.6)4.130,446 (2,181)7.2
Flow cytometry and compensation

The compensation matrices for each combination of fluorochrome-conjugated antibodies were performed as above compared.

Compensation Matrices for Different Instruments

Monoclonal antibodies

CD3-PE (UCHT1), CD3-PC5.5 (UCHT1), CD3-APC-Alexa750 (UCHT1), CD3-PB (UCHT1), CD5-PC7 (SFCI12T4D11), CD5-APC (SFCI12T4D11), CD5-APC-Alexa750 (SFCI12T4D11), CD8-PC5.5 (B9.11), CD8-KO (B9.11), CD16-PB (3G8), CD19-PC5.5 (J3 119), CD19-APC (J3 119), CD20-APCAlexa-750 (89B), CD36-FITC (FA6.152), CD41-PC7 (P2), CD42b-FITC (SZ2), CD45-FITC (J.33), CD45-PE (J.33), CD45-PC5.5 (J.33), CD45-PC7 (J.33), CD45-APC (J.33), CD45-APC (J.33), CD45-APC-Alexa750 (J33), CD45-PB (J.33), CD45-KO (J.33), CD56-PE ((N901), CD79a-APC (HM47), CD117-PC7 (95C3), CD160-PE (BY55), and FMC7-FITC (FMC7) were from Beckman Coulter.

Method

Four combinations of eight antibodies were selected, each containing antibodies conjugated to FITC, PE, PC5.5, PC7, APC, APC-Alexa Fluor750 (APC-AF750), Pacific Blue (PB), and Krome Orange (KO) (Supporting Information Table 2). The combinations were then used to label CompBeads according to the manufacturer's instructions.

Table 2. Compensation Matrices for Tandem and Non-Tandem mAbs [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
 Percent spillover into:
 mAbFITCPEPerCP-Cy5.5PE-Cy7APCAPC-Cy7V500
  1. The spillover correction factors for mAbs within a given non-tandem group of mAbs are very similar. This was the case also for the PE-Cy7-conjugated mAbs tested.

Monoclonal antibodies conjugated to:FITCCD57 15.262.380.250.000.003.45
CD11b 15.362.490.310.000.003.79
MPO 15.182.430.330.020.023.88
CD2 14.642.250.290.000.013.80
CD103 15.142.420.280.000.003.72
CD81 15.272.500.300.000.013.98
 CD38 15.852.570.280.060.004.00
PECD131.10 18.022.000.020.000.01
CD561.10 18.172.020.010.000.02
CD79a1.15 18.312.010.040.010.02
CD1601.15 18.232.070.030.000.05
CD1231.11 18.142.030.020.000.01
CD141.11 18.042.020.020.000.02
 CD81.05 17.662.280.040.000.07
PE-Cy7CD1170.101.713.25 0.031.480.00
CD50.101.733.28 0.031.490.03
CD100.091.253.01 0.021.550.03
CD250.111.763.39 0.061.410.02
CD190.091.503.04 0.021.690.01
CD380.101.503.09 0.021.510.00
 CD140.101.653.00 0.041.530.04
APCCD330.010.001.090.17 2.840.00
CD220.000.001.090.16 3.500.05
CD230.010.001.100.17 2.850.02
CD50.020.011.080.16 2.870.04
CD11c0.010.001.060.18 2.870.01
CD1380.010.001.090.18 2.880.02
 CD550.000.001.090.19 3.100.05
APC-Cy7CD30.070.000.857.73 31.070.13
CD450.090.020.618.00 30.570.05
 CD200.150.050.678.64 20.260.20
Flow cytometry

The compensation matrices for each combination of antibodies were determined using software available on the two instruments: A Canto II with standard laser and filter configuration and FACSDiva software version 6.1.3 and a Navios (Beckman Coulter) with a standard three color configuration and software. Both instruments were optimized for use and the optimized PMT voltages were used for the compensation protocol. The compensation matrices for each combination of fluorochrome-conjugated antibodies were compared on the two instruments.

Spectral Spillover of Commonly Used Tandems from Different Manufactures and Spectral Spillover of Violet-Excited Fluorochromes

Monoclonal antibodies

CD3-APC-H7 (SK7), CD3-PB (UCHT1), CD4-V500 (RPAT4), CD5-PE-Cy7 (LI7F12), CD4-BV421 (RPA-T4), CD45-V500 (2D1), and CD56-V450 (B159) were from BD Biosciences. CD3-APC-AF700 (UCHT1), CD5-PC7 (BL1a), CD20-PB (B9E9), and CD45-KO (J.33) was from Beckman Coulter. CD3-APC-eFluor750 (UCHT1) and CD5-PE-Cy7 (UCHT2) was from eBioscience.

Method

Compensation matrices for tandems and violet laser-excited fluorochromes from three or two different manufacturers, respectively, were performed on the same day, and in the same FACSDiva software version 6.1.3 compensation set-up experiment.

Flow cytometry and compensation set-up

As described above.

Lot-To-Lot Variation of Spectral Spillover

Monoclonal antibodies

CD5-PE-Cy7 (LI7F12), CD3-APC-H7 (SK7), CD117-PE-Cy7 (104D2), and HLA-DR-APC-H7 (L243), all from BD Biosciences.

Method

Lot-to-lot comparisons were performed on separate days in two laboratories. New lots were compared with old lots by over-acquiring the old tandem conjugate data in the FACSDiva software version 6.1.3 compensation set-up experiment. Since the same compensation set-up experiment was used for all lots tested, the PMT voltages used were constant over time: PMT voltage adjustments for day-to-day drift were not applied. The average CS&T delta voltage adjustment requirement during the time period concerned did not exceed 20 V for any channel.

Flow cytometry and compensation set-up

As described above.

Interpretation of Incorrectly Compensated Data

Monoclonal antibodies

CD38-PE (HB7), CD26-PE (L272), CD5-PerCP-Cy5.5 (L17F12), and CD3-APC-H7 (SK7) were from BD Biosciences; CD5-PC7 (BL1a), CD10-PC7 (ALB1), and CD19-PC7 (J3–119) were from Beckman Coulter, CD7-APC (eBio124-1D1) was from eBioscience.

Method

FCS data files derived from routine diagnostic 7 or 8 color screens were analyzed in FACSDiva software version 6.1.3. Using the correct compensation matrix as a starting point, the matrix related to tandem spillover was increased or decreased in a step-wise manner using off-line compensation. The visual interpretation of antigen expression patterns, median fluorescence intensity (MFI) and population size (% positive cells) of gated populations were measured for each compensation setting.

Routine diagnostic specimens were labeled with antibodies for 15 minutes at room temperature in the dark. After red cell lysis (Pharmlyse, BD Biosciences, used according to manufacturer's instructions), samples are washed in PBS and centrifuged at 600g for 4 minutes. Samples were acquired within 30 minutes.

Flow cytometry

As described above. The cytometer was set at medium flow rate; acquisition rate did not exceed 8,000 events/second. The window extension was 7.0, forward and SSC voltages and FSC threshold were set to include all leukocyte populations, including a complete lymphocyte cloud and, in some cases, the complete nuclear red cell cloud.

Statistical Analysis

For the stability of cocktails and the variation in compensation matrices, the mean, standard deviation (SD) and coefficient of variation (CV) were calculated using Microsoft XL software.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. LITERATURE CITED
  9. Supporting Information

Stability of Tandems in Antibody Cocktails Used to Label Preserved Blood and Fresh Bone Marrow

Preserved blood

On initial analysis, the CytoComp cells bound all cocktail antibodies except CD56-V450, Kappa-FITC, and Lambda-PE. The Multi-Check Control bound all antibodies. Table 1 shows the mean values for the percentage positive cells and median fluorescence intensities of the eight analyses performed on CytoComp cells. There was very little variation in the percentage positive values, with CVs of five or less for the six fluorochromes analyzed. The CVs for the MFI's were higher, though all were below 10. Similarly, the results for the six analyses of the Multi-Check Control showed very little variation in the percentage positive values, with CVs of five or less. The exception was CD56-V450, which had a CV of 12.3. There was more variability in MFI's with CV < 10 for all antibodies tested except CD56-V450, which had a MFI CV of 23.7 (Table 1). The only consistent change over time was a reduction in CD8-PE MFI on the Multi-Check control (Supporting Information Table 3). This, however, was not seen during the eight weeks testing of the Cytocomp cells (Supporting Information Table 4). Visual interpretation of expression patterns did not give cause for any concerns; and neither cell type had any evident increase in background fluorescence over time for any of the fluorochromes used (Fig. 1).

Table 3. Comparison of Spectral Overlaps for PE-Cy7 Conjugates from Different Manufactures
PercentPE-Cy7/PC7
eBioscience (UCHT2)Beckton Dickinson (LI7F12)Beckman Coulter (BL1a)
  1. The mAbs tested were different CD5-specific clones.

FITC0.070.050.07
PE1.241.650.93
PerCP-Cy5.53.463.683.24
APC0.020.050.04
APC-H72.632.422.91
Pacific Blue0.000.000.00
V5000.000.000.00
Table 4. Variation in Spectral Overlap Between Different APC-Tandem Conjugates
PercentAPC-Tandem
APC-eFluor750 (UCHT1)APC-H7 (SK7)APC-AF700 (UCHT1)
  1. The antibodies were all specific for CD3, the clone and tandem conjugates are indicated in the table.

FITC0.000.000.00
PE0.000.000.00
PerCP-Cy5.50.560.230.48
PC75.546.385.24
APC49.8011.4040.68
Pacific Blue0.000.000.00
V5000.000.000.00
image

Figure 1.  CytoCheck cells labeled with the cocktail of antibodies. A Cells labeled with cocktail at beginning of stability study. B Cells labeled with the cocktail after eight weeks of storage. Very little change in expression patterns or background fluorescence was observed. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Download figure to PowerPoint

Fresh bone marrow

MFIs and population sizes for all antigens and populations investigated remained very similar between cells labeled with cocktail or daily individually aliquoted mAbs, over the 5-day test period. A slight decrease in HLA-DR-PB on monocytes labeled with the cocktail, compared to daily individually aliquoted mAbs, was found. This difference did not influence the proportions of the HLA-DRdim-mod CD11bdim mature/HLA-DRbright CD11bdim immature monocytes as determined by the HLA-DR/CD11b gate. The CD16 MFI on mature neutrophils was marginally higher for samples labeled with the cocktail (Supporting Information Table 5). Again, this did not affect the maturation patterns studied, which all rely on the range of CD11b, CD13, and CD16 expression intensities on the maturing cells. There was no obvious difference in background fluorescence between cocktail and individually aliquoted mAbs at any time point tested. In particular, looking at signals from the tandem donor fluorochromes, we found no increase in CD33-APC signals of either APC positive populations (CD33+ Monocytes), APC dim populations (CD33+ Neutrophils), or fully negative cells (CD33 lymphocytes) (Supporting Information Table 5 and data not shown). Similarly, cocktail generated PE signals from CD13-PE dim immature neutrophils or CD13 lymphocytes did not increase over time (Supporting Information Table 5 and data not shown).

Table 5. Spectral Overlap for Antibodies Conjugated to 420–455 nm Emitting Fluorochromes
PercentCD20-PBCD3-PBCD19-V450CD56-V450CD4-BV421
FITC1.791.871.551.541.18
PE0.160.230.150.190.12
PerCP-Cy5.50.000.010.030.030.00
PC70.000.000.000.000.00
APC0.020.000.050.000.01
APC-H70.000.010.000.000.00
KO45.5545.8338.7638.3116.54

Finally, data were examined for artifacts, which were defined as additional populations seen in the cocktail-labeled samples, but not detected in the individually aliquoted tests. No artifacts were identified in any of the cocktail-labeled samples upon close examination systematically plotting each fluorochrome against all others (data not shown).

Comparison of Compensation Matrices for mAbs Conjugated to the Same Fluorochrome

Most clinical laboratories use a generic compensation matrix for all antisera conjugated to non-tandem fluorochromes. For example, all FITC-conjugated mAbs would use the same matrix. To explore whether a generic compensation matrix could be used also for tandem-conjugated mAbs, the matrices for seven separate FITC-, PE-, APC-, and PE-Cy7-conjugated mAbs and three different APC-Cy7 mAbs were compared. The largest variation within a group was for the APC-Cy7-conjugated mAbs: their APC-Cy7 spillover into the APC channel varied from 20.26% to 31.07% (Table 2). The correction factors for all PE-Cy7-conjugated mAbs were quite similar, and did not markedly differ more than those for non-tandem-conjugated mAbs (Table 2).

Compensation Matrices for Different Instruments

The four antibody cocktail combinations are shown in Supporting Information Table 2. The Mean and SD for the compensation matrices are shown in Supporting Information Figure 1. The compensation matrices were similar on the two instruments.

Spectral Overlap Between Tandems from Different Manufacturers

Compensation matrices for PE-Cy7/PC7-conjugated mAbs from three separate manufacturers are shown in Table 3. The greatest difference was for spillover into the PE channel: The difference between the maximum and minimum values in absolute terms was 0.72%. APC-tandems that use different receiver fluorochromes were also tested. Their correction factors for APC varied greatly: the maximum difference was 38.4%. Their PC7 correction factors varied from 5.24% to 6.38% (Table 4).

Spectral Overlap for Violet-Excited Fluorochromes

Tables 5 and 6 show compensation matrices for violet-excited fluorochromes from two different companies. The ‘blue' emitting fluorochromes PB, V450 and BV421 mainly spillover into the ‘yellow' 510 nm ± 25 nm channel. Notably, this was less for BV421 than for either PB or V450 (Table 5). KO and V500 mainly spillover into the ‘blue' 450 nm ± 25 nm channel and to a lesser extent, into the FITC channel (Table 6).

Table 6. Spectral Overlap for Antibodies Conjugated to 500–530 nm Emitting Fluorochromes
PercentV500/KO
CD45-V500CD4-V500CD45-KO
FITC1.341.450.63
PE0.030.340.06
PerCP-Cy5.50.020.050.00
PC70.020.000.10
APC0.000.000.00
APC-H70.020.000.03
Pacific Blue7.777.371.22

Correction Factor Lot-To-Lot Variation for Tandem mAbs

Compensation matrices for five different lots of the same CD5-PE-Cy7 or CD3-APC-H7 mAb, recorded on the same cytometer in laboratory 1 are shown in Figures 2A and 2B. Similarly the compensation matrices for five different lots of the same CD117-PE-Cy7 or HLA-DR-APC-H7 mAb, recorded on the same cytometer in laboratory 2 are shown in Figures 2C and 2D. The two laboratories produced very similar compensation matrices for the conjugates even though different monoclonal antibodies were studied. The variation in the % compensation was less than 2% for all conjugates investigated. Several lots of other tandem-conjugated mAbs, including PC7 and PerCp-Cy5.5 conjugates, were tested in the same way by laboratory 1 and 2 and were found to have less variation than CD5-PE-Cy7 or CD3-APC-H7 (data not shown).

image

Figure 2.  Lot-to-lot spillover variation for tandem mAbs. The spectral overlap for five different lots each of different tandem mAbs were collated in two separate laboratories. A CD5-PE-Cy7, laboratory 1; B CD3-APC-H7, laboratory 1; C CD117-PE-Cy7, laboratory 2, and D HLA-DR-APC-H7, laboratory 2. The largest correction factor difference between the APC-H7 mAb slots was 1.75%, for spillover into APC. For the PE-Cy7 mAbs lots, the largest variation was 1.84%, for spillover into APC-H7.

Download figure to PowerPoint

Interpretation of Incorrectly Compensated Data

Please see discussion below.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. LITERATURE CITED
  9. Supporting Information

Tandem Conjugate Stability in Antibody Cocktails

The main reason for false results in multicolor flow cytometry is thought to be incorrect compensation. This is not disputed here; yet, the flow cytometry-related non-conformities that have occurred since the introduction of MCF in our laboratories are chiefly related to the addition of incorrect antisera to sample tubes or the omission of an antibody. Notably, these incidences occur more frequently for panels that do not have pre-made cocktails. The use of cocktails is, in our experience, helpful as it saves time and avoids human error. However, to date, few publications have reported the stability of antisera, including tandem mAbs, used in cocktails. Biancotto et al. [1] tested a cocktail containing 14 mAbs over a 4-day period, and found that the qdot 605-, qdot 655-, and PE-Cy5-conjugated mAbs did not perform well after 3 days. Interestingly, the authors did not report a problem with the other tandem mAbs used in the cocktail: APC-Cy7, PE-Cy7, PerCP-Cy5.5, PE-TR and PE-Cy5.5-conjugated mAbs. Rawstron et al. [2] investigated two separate cocktails, each containing six mAbs (including APC-H7- and PE-Cy7-conjugated mAbs), over a 28-day period. They observed no loss of the PE-Cy7 or APC-H7 signals during this time, but did detect an increase in background APC-H7 intensity on APC positive cells. This effect was limited to one of the cocktails, and to cells with strong expression of the antigen binding the APC-conjugated mAb.

The present study tested two cocktails. Both cocktails contained antibodies from different manufacturers and therefore possibly a mixture of different buffers. The first cocktail was used to examine lymphocyte subsets and was tested on commercially available preserved and lyophilized cell controls over a 6 and 8-week period, respectively. A consistent decrease in CD8-PE MFI was observed for the preserved cell control but not for the lyophilized cell control. This would suggest that the cocktail used performed equally well at all time points, while the preserved cells had reduced capacity for binding the CD8 antibody over time. Possibly, freshly reconstituted lyophilized cells offer a more stable vehicle for testing antibody cocktails. The second cocktail targeted myeloid antigens and was tested on fresh bone marrow samples over a 5-day period. For both the cocktails, none of the observed changes would have altered the clinical outcome: Visual interpretation of expression patterns did not reveal differences, nor did the minor numerical differences influence data interpretation. The data presented here, along with that reported by Biancotto et al. [1] and Rawstron et al. [2], suggest that the performance of each separate cocktail needs to be evaluated, and that performance may differ depending on the test material. Here, we used three substrates for testing the cocktails: preserved or lyophilized control cells and fresh bone marrow samples. Preserved and lyophilized cells have the advantage of consistency: the population size and fluorescence intensity should remain the same throughout the test period. Fresh samples include all normal cell populations and therefore allow for monitoring non-specific binding of several cell subsets as well as performance of antibodies specific for antigens with different expression levels on the same, or different, cell subsets. The possibility of tube-to-tube variation, rather than cocktail shelf-life, causing different results may be greater when using fresh samples. It should also be noted that none of these choices offers the possibility of testing cocktails on neoplastic populations. Le Roy et al. [14] showed that APC-tandems may be degraded depending on cell type present in the sample. The effect was relatively immediate (within 30 minutes) and also occurred in single mAb-stained samples. It is, therefore, difficult to see how results influenced by cell type in the manner described by Le Roy et al. [14] would be different should the APC tandem be stored singly or in a cocktail.

PE-tandem mAbs, particularly PE-Cy7, are photosensitive [7]. Both cocktails tested here were stored in light protective bottles, and were subjected to a single daily light exposure of 10 minutes (when used to label preserved cells) or 20 minutes (when used to label fresh bone marrow). The different exposure levels were selected because they reflect the average time spent on the bench for the respective cocktails in our two laboratories. This level of light exposure did not appear to cause tandem degradation, at least not at a level that affected the data analyzed here.

Comparison of Compensation Matrices for mAbs Conjugated to the Same Fluorochrome

Clinical labs in the UK often use a generic antibody when setting compensation matrices for non-tandem fluorochromes. Analysis of groups of 7 different FITC-, PE-, APC-, and PE-Cy7-conjugated mAbs showed that the PE-Cy7 spillover into the PE channel did not vary over that of, for example, APC spillover into the APC-Cy7 channel. Most variation in the required compensation occurred for the APC-Cy7-conjugated mAbs, between APC and APC-Cy7. This may be due to the variation in protein:fluorochrome ratio for these tandems or the efficiency of the resonance energy transfer between the two fluorochromes in that particular tandem conjugate. It is however difficult to compare spillover values for the different fluorochromes. The degree of spillover is diverse, ranging from an approximate 15 % for FITC into PE to below 0.1% for other combinations. Moreover, the effect of correction factor change on the cell population MFIs for the different fluorochromes may not be the same across all combinations.

Compensation Matrices for Different Instruments

Comparison of compensation matrices on Canto II and Navios showed that the greatest amount of compensation was required for PE – FITC, PC7 – PC5.5, APC – APC-Alexa750 and KO – PB on both instruments. The variations observed between the two cytometer's matrices reflect their different filter set up, and would not prevent standardization and harmonization on different types of instruments.

Violet-Excited Fluorochromes

It is generally accepted that non-tandem fluorochromes have very similar spillover characteristics regardless of production line and most clinical laboratories would use them interchangeably in compensation matrices. Newer organic fluorochromes that are excited by violet lasers are now used extensively by many laboratories. These include dyes such as PB, V450, brilliant blue, and brilliant violet dyes [15], normally detected using 450 nm ± 25 nm filters. Also KO and V500 that are normally detected using 510 nm ± 25 nm filters. So far, our experience is that PB, V450, V500, and KO do not vary more than other organic dyes from lot-to-lot. However, more extensive data sets should be recorded to confirm this observation. Despite fairly similar emission spectra, the spillover corrections required for these fluorochromes are different, as shown in Tables 5 and 6. Off-line compensation checks showed that their compensation matrices cannot be used interchangeably in our set-up (data not shown). To the contrary, the two main different ‘green'-emitting fluorochromes used in clinical laboratories, FITC, and Alexa 480, have very similar matrices in our systems and may be used interchangeably (data not shown). Our data highlights that individual matrices are required for the different violet laser exited fluorochromes.

Correction Factor Variation Between Lot-To-Lot and Different Manufacturers

Due to the lot-to-lot variation of tandem conjugates we investigated whether different compensation values may be required for each lot. Different antibodies conjugated to PE-Cy7 or APC-H7 were found to have relatively similar compensation matrices for different lots. In addition, the compensation matrices for the tandems were found to be very similar even when analyzed on different instrument from the same manufacturer in different laboratories. The difference in compensation between lots was always <2% and in most cases <1%. Some of the differences between the matrices shown in Figure 2 are likely due to daily drift, since PMT voltages were not corrected for day-to-day variations in cytometer performance (see “MATERIALS AND METHODS”). The matrices for PE-Cy7-conjugated mAbs from three different companies did not vary over those for the different PE-Cy7 lots from one single company. It should be noted that some of the companies may use the same production line. In contrast, the APC-tandems tested varied greatly, particularly for spillover into the APC channel. This was expected, since many companies use different receiver fluorochromes for their APC-tandems, for example, H7, Cy7, and Alexa Fluor 700/750, which was the case for the mAbs tested here.

Interpretation of Incorrectly Compensated Data

We have shown that the spectral spillover for different lots of tandem antibodies may vary. This raise the question whether the variation observed would affect data interpretation? The extent of the effect of incorrect compensation on MCF data depends not only on the magnitude of the change in the correction factor, but also on the particular antibody/fluorochrome combinations used, and on the expression patterns studied. We therefore investigated several FCS files derived from various antibody and cell combinations, using correct and incorrect off-line compensation. Importantly, numerical evaluation of results is not always the unique way to check changes since visual interpretation based on pattern comparison plays a large role in diagnostic procedures. Therefore, expression patterns, as well as percentage positive cells and MFIs for gated populations were evaluated.

The greatest effect tends to be observed for antigens that are weakly expressed where the population of interest ranges from a negative to weakly/moderate pattern of antigen expression (Supporting Information Figs. 2b and 2c) when compared to more highly expressed antigen (Supporting Information Fig. 2a). Where the population of interest is relatively well separated from the remaining cells, as is the case for the CD7 negative T cells shown in Supporting Information Figure 2d, even larger deviations from correct compensation settings (up to 10%) would not cause a false result. However, should a population be defined as CD5+CD26−/dim (as in Supporting Information Fig. 2b), the reported population size would be affected by compensation changes of <2%. That said, a population is rarely defined by one antigen alone but, rather, by a series of gates (sequential Boolean gating). For the example shown in Supporting Information Figure 2b, the Sezary cells were also identified by the complete loss of CD7, and the population size was gated on according to CD3+/CD4+/CD7/CD26−/dim expression. Similarly, the CD5+ B cells shown in Supporting Information Figure 2c would be investigated for expression of several other antigens before reported as “abnormal”. In addition, in many cases the analyst would notice that the data was either over or under compensated.

Perhaps of most concern is the effect of incorrect compensation on small population sizes that are defined by altered antigen expression, rather than by a clear positive or negative population. An example is MRD analysis. Using data files for B-ALL and AML MRD-positive and -negative cases, it was not possible to create a false-positive or -negative result by changing one or several correction factors within 2%, including channels used by non-tandem antibodies (data not shown). Similarly, it was not possible to alter the interpretation of MDS-related monocyte and neutrophil maturation patterns [16]. All these tests were based on populations defined by four antigens, and the analysis was always interpreted manually. Current and, particularly, future computer-based analysis may take smaller changes in fluorescence intensity into account than are currently used for manual analysis.

What is A Clinically Safe Compensation Strategy for Busy Laboratories?

The safest strategy for obtaining accurate data must surely be to check the spillover characteristics for each new lot of tandem antibody. If the lot is different from the previous lot, then compensation must be performed. It is not possible to advocate the creation of an inexact system or approach to diagnostic MCF. At the same time, it is recognized that many clinical laboratories have time pressures and, sometimes, a lack of staff with adequate training in cytometry. Thus, it is realistic to assume that, at least on occasion, re-compensation of new tandem antibody lots has not taken place prior to data reporting, and/or that compensation settings for one PE-Cy7-conjugated mAb are used for a different PE-Cy7-conjugated mAb.

Knowing the spillover variation between tandem lots and between tandem mAbs from different companies is helpful for evaluating how the final result might be influenced using a wrong compensation matrix. We have shown here that for at least certain tandem mAbs, compensation for each new lot is not needed, and it could also be safe with a generic compensation matrix for the PE-Cy7 mAbs used in this study. It must be emphasized that each tandem and the combinations it is used in require validation prior to using such protocols. Further strategies that guard against false results derived from incorrect compensation would include choosing the appropriate hardware, such as lasers, mirrors and filters, as well as careful selection of fluorochrome conjugates. If possible, the same configuration and voltage settings could be used for all panels and tubes, since the application of a wrong compensation matrix would then cause less erroneous analysis. It would also be useful to identify antibody combinations that are more susceptible to smaller (up to 2%) compensation deviations.

This study focused on a few commonly used tandems. Several tandem fluorochromes are currently available, and more are likely to be developed. It is hoped that these, and other fluorochromes, will enable further expansion of clinical MCF. The tests carried out in this study would be useful for selecting which fluorochromes and dyes to include in assays that require a robust performance.

Looking to the future, manufacturers may provide lyophilized cocktails. These could be made from large batches of mAb and validated by the manufacturer such that lot-to-lot variation is reduced. This would alleviate some of the consideration commented upon here.

ACKNOWLEDGMENTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. LITERATURE CITED
  9. Supporting Information

The authors would like to thank Nan Jiang and Kelly Lundsten at Biolegend for useful discussions and Nicki Senior for help with the compensation matrix analysis performed on the Navios.

LITERATURE CITED

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. LITERATURE CITED
  9. Supporting Information
  • 1
    Biancotto A, Fuchs JC, Williams A, Dagur PK, McCoy JP Jr. High dimensional flow cytometry for comprehensive leukocyte immunophenotyping (CLIP) in translational research. J Immunol Methods 2011;363:245261.
  • 2
    Rawstron AC, Bottcher S, Letestu R, Villamor N, Fazi C, Kartsios H, de Tute RM, Shingles J, Ritgen M, Moreno C, et al. Improving efficiency and sensitivity: European Research Initiative in CLL (ERIC) update on the international harmonised approach for flow cytometric residual disease monitoring in CLL. Leukemia 2013;27:142149.
  • 3
    Forman MA, Gupta RK. Tandem dyes for flow cytometry: Can we overcome quality concerns? MLO Med Lab Obs 2007;39:24, 26.
  • 4
    Stelzer GT, Marti G, Hurley A, McCoy P Jr, Lovett EJ, Schwartz A. U.S.-Canadian Consensus recommendations on the immunophenotypic analysis of hematologic neoplasia by flow cytometry: Standardization and validation of laboratory procedures. Cytometry 1997;30:214230.
  • 5
    Stewart CC, Stewart SJ. Four color compensation. Cytometry 1999;38:161175.
  • 6
    Baumgarth N, Roederer M. A practical approach to multicolor flow cytometry for immunophenotyping. J Immunol Methods 2000;243:7797.
  • 7
    Hulspas R, Dombkowski D, Preffer F, Douglas D, Kildew-Shah B, Gilbert J. Flow cytometry and the stability of phycoerythrin-tandem dye conjugates. Cytometry A 2009;75A:966972.
  • 8
    Wood B. 9-color and 10-color flow cytometry in the clinical laboratory. Arch Pathol Lab Med 2006;130:680690.
  • 9
    Matarraz S, Lopez A, Barrena S, Fernandez C, Jensen E, Flores-Montero J, Rasillo A, Sayagues JM, Sanchez ML, Barcena P, et al. Bone marrow cells from myelodysplastic syndromes show altered immunophenotypic profiles that may contribute to the diagnosis and prognostic stratification of the disease: A pilot study on a series of 56 patients. Cytometry B Clin Cytom 2010;78B:154168.
  • 10
    van Lochem EG, van der Velden VH, Wind HK, te Marvelde JG, Westerdaal NA, van Dongen JJ. Immunophenotypic differentiation patterns of normal hematopoiesis in human bone marrow: Reference patterns for age-related changes and disease-induced shifts. Cytometry B Clin Cytom 2004;60B:113.
  • 11
    Wood BL. Flow cytometric diagnosis of myelodysplasia and myeloproliferative disorders. J Biol Regul Homeost Agents 2004;18:141145.
  • 12
    Wells DA, Benesch M, Loken MR, Vallejo C, Myerson D, Leisenring WM, Deeg HJ. Myeloid and monocytic dyspoiesis as determined by flow cytometric scoring in myelodysplastic syndrome correlates with the IPSS and with outcome after hematopoietic stem cell transplantation. Blood 2003;102:394403.
  • 13
    Arnoulet C, Bene MC, Durrieu F, Feuillard J, Fossat C, Husson B, Jouault H, Maynadie M, Lacombe F. Four- and five-color flow cytometry analysis of leukocyte differentiation pathways in normal bone marrow: A reference document based on a systematic approach by the GTLLF and GEIL. Cytometry B Clin Cytom 2010;78B:410.
  • 14
    Le Roy C, Varin-Blank N, Ajchenbaum-Cymbalista F, Letestu R. Flow cytometry APC-tandem dyes are degraded through a cell-dependent mechanism. Cytometry A 2009;75A:882890.
  • 15
    Chattopadhyay PK, Gaylord B, Palmer A, Jiang N, Raven MA, Lewis G, Reuter MA, Nur-ur Rahman AK, Price DA, Betts MR, et al. Brilliant violet fluorophores: A new class of ultrabright fluorescent compounds for immunofluorescence experiments. Cytometry A 2012;81A:456466.
  • 16
    Westers TM, Ireland R, Kern W, Alhan C, Balleisen JS, Bettelheim P, Burbury K, Cullen M, Cutler JA, Della Porta MG, et al. Standardization of flow cytometry in myelodysplastic syndromes: A report from an international consortium and the European LeukemiaNet Working Group. Leukemia 2012;26:17301741.

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. LITERATURE CITED
  9. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
cytob21154-sup-0001-suppfig1.pptx120KSupplementary Information Figure 1
cytob21154-sup-0002-suppfig2.pptx1883KSupplementary Information Figure 2
cytob21154-sup-0003-supptables.docx24KSupplementary Information Tables
cytob21154-sup-0004-suppinfo.docx14KSupplementary Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.