Discrepancy in measuring CD4 expression on T-lymphocytes using fluorescein conjugates in comparison with unimolar CD4-phycoerythrin conjugates

Discrepancy in Measuring CD4 Expressionon T-Lymphocytes Using Fluorescein

Conjugates in Comparison with UnimolarCD4-Phycoerythrin Conjugates

Lili Wang,1* Fatima Abbasi,2 Adolfas K. Gaigalas,1 Robert A. Hoffman,3

Dan Flagler,4 and G. E. Marti21National Institute of Standards and Technology (NIST), Gaithersburg, Maryland 20899-8312

2Center for Biologics Evaluation and Research, U.S. Food and Drug Administration, Bethesda, Maryland 208923BD Biosciences, 2350 Qume Drive, San Jose, California 95131

4Beckman Coulter Inc., PO Box 169015, Miami, Florida 33116-9015

Background: Numerous methods for quantitative fluorescence calibration (QFC) have been developed toquantify receptor expression on lymphocytes. However, the results from the use of these different QFCmethods vary considerably in the literature. To better identify the causes of these discrepancies, wemeasured CD4 expression using FITC and phycoerythrin (PE) conjugates to stain CYTO-TROLTM ControlCells and T-lymphocytes in whole blood and isolated cell preparations. We further examined pH of thecellular microenvironment as a cause of discordant results obtained with the FITC conjugate.Methods: Calibration with Quantibrite PE-labeled microspheres and the use of unimolar CD4-PE conju-

gates provided direct measurement of the antibody bound per cell value (ABC) for CD4 expression on nor-mal T-lymphocytes. Calibration for CD4-FITC monoclonal antibody (Mab) labeled CYTO-TROL Control Cellsand normal T-lymphocytes was based on molecules of equivalent soluble fluorochrome (MESF) as deter-mined by FITC-labeled microspheres traceable to NIST RM 8640. The MESF value for CD4-FITC Mab wasdetermined that enabled the conversion of the MESF values obtained for CYTO-TROL cells to ABC. Weinvestigated the likely pH change in the fluorescein microenvironments within FITC-labeled Mab andcells stained with FITC-labeled Mab using a pH sensitive indicator.Results: The mean ABC value for T-lymphocytes prepared from fresh whole blood using CD4-PE conju-

gate (48,321) was consistent with previous results, and it was much higher than the mean ABC usingCD4-FITC Mab (22,156). The mean ABC value for CYTO-TROL cells using CD4-PE conjugate (43,090) wasalso higher than that using CD4-FITC conjugate (34,734), although the discrepancy was not as great. Fur-ther studies suggested the discrepancy in CYTO-TROL results may be accounted for by the low pH of themembrane microenvironment, but the greater discrepancy in T-lymphocytes could not be fully explained.Conclusion: CD4 expression on fresh normal whole blood samples and CYTO-TROL cells can be consis-

tently quantified in ABC units using Quantibrite PE quantification beads and unimolar CD4-PE conjugates.Quantification with CD4-FITC conjugate is not as consistent, but may be improved by the use of CD4 T-cellsas biological calibrators. This approximation is valid only for surface receptors with consensus ABC valuesmeasured by different QFC methods serving as biological standards. Published 2007 Wiley-Liss, Inc.{

Key terms: CD4; QuantibriteTM PE quantification kits; ABC value; MESF value; FITC-labeled microspheres;CYTO-TROL Control Cells; pH effect; antibody binding affinity; effective F/P ratio

{This article is a US government work and, as such, is in the publicdomain in the United States of America.

This work does not represent an official position of the Food andDrug Administration.

*Correspondence to: Lili Wang, National Institute of Standards andTechnology, 100 Bureau Drive, Stop 8312, Gaithersburg, MD 20899-8312, USA. E-mail: [email protected] or Gerald E. Marti, Center forBiologics Evaluation and Research, U.S. Food and Drug Administra-

tion, NIH Building 29B, Room 2NN08, Bethesda, MD 20892, USA.E-mail: [email protected] 13 December 2006; Revised 2 March 2007; Accepted 15

March 2007Published online 1 May 2007 in Wiley InterScience (www.

interscience.wiley.com).DOI: 10.1002/cyto.b.20354

Cytometry Part B (Clinical Cytometry) 72B:442–449 (2007)

Published 2007 Wiley-Liss, Inc.{

The use of quantitative immunofluorescence measure-ments has increased significantly since the mid-1990s.Numerous methods for quantitative fluorescence calibra-tion (QFC) have been developed to quantify receptorexpression on a cell. Though the ultimate goal of quanti-fication is to measure the number of antigens or ligandbinding sites on a cell, this task is carried out by measuringthe number of antibodies bound per cell (ABC). In 1998, aspecial issue of the journal Cytometry was dedicated tothe topic of ‘‘Quantitative Fluorescence Cytometry: anEmerging Consensus’’ (1) which provided a comprehen-sive status report on the quantification effort. There was aconsensus that fluorescence signal should be quantified interms of molecules of equivalent soluble fluorochrome(MESF). It was also generally agreed in the community thatthe current goal is to develop practical, reliable, and yetinter-laboratory comparable methods for quantifying thenumber of fluorescent antibodies or other biomoleculesbound on cells.

Five QFC methods have been developed and used forquantification including MESF FITC microsphere calibra-tion, the use of Quantibrite calibration beads, quantita-tive indirect immunofluorescence assay (QIFI), quantumsimply cellular (QSC) bead method, and the use of CD4on normal lymphocytes as a biological calibrator (2,3).The first two methods are based upon the MESF con-cept. The assignment of MESF values relies on the equal-ity of fluorescence yields from two solutions defined asthe product of fluorochrome concentration and the mo-lecular quantum yield (4). With the use of a set ofmicrobeads, each with a known MESF value, the MESFvalue of an analyte of interest can be determined on thebasis of a flow cytometric calibration curve generated byusing the reference bead set (5). To obtain the ABCvalue of the monoclonal antibody (Mab) specific for theanalyte, the MESF value or the effective fluorophore perantibody (F/P) ratio of the Mab is required. The ABCvalue is the ratio between the MESF value of the cellsstained with labeled Mab and the MESF value of the la-beled antibody. If the manufacturer could provide theMESF value or the effective F/P ratio for the conjugatedMab, it would be straightforward for users to calculatethe ABC value for the conjugated Mab. This value is notavailable for most Mab, such as fluorescein-labeled Mab.Hence, it is not yet practical for the determination ofABC values with the use of fluorescein as the labelingfluorochrome. Quantibrite calibration beads (BD Bio-science), on the other hand, are a set of microbeadswith different amount of immobilized phycoerythrin(PE) molecules. By using unimolar conjugates of PE toMab (effective F/P ratio ¼ 1), ABC values of PE-labeledMab can be determined directly according to the calibra-tion curves generated by utilizing Quantibrite beads.

Quantitative indirect immunofluorescence assay, QIFI,was first developed by Poncelet and Carayon (5). Thismethod employs a series of microbeads coated with dif-ferent known amounts of mouse immunoglobulin in anattempt to mimic cells coated with Mabs at saturatingconcentration. A second layer, composed of labeled goat

anti-mouse IgG under saturating conditions, is then usedto calibrate flow cytometers in terms of antibody bindingcapacity. By measuring fluorescence signals from biologi-cal cells incubated with saturating amounts of Mabs, fol-lowed by incubation with labeled secondary antibodies,ABC values can be determined from the calibrationcurves. Using the QIFI method, however, Bikoue et al.observed that different Mabs recognizing the same anti-gen could give different ABC values on the same cells(6). Keep in mind that different Mabs might recognizedifferent epitopes of the same receptor and Mab bindingcould be monovalent or bivalent depending on theepitope (7).The QSC bead method utilizes a series of four micro-

sphere populations labeled with different amounts ofgoat anti-mouse or goat anti-human IgG. These micro-spheres are stained in parallel with the same fluores-cently labeled Mab as cells. Each microsphere populationwill bind a known amount of Mab and hence, four popu-lations produce a cytometric calibration curve for meas-uring the ABC value of cells. Similar to the QIFI methoddescribed earlier, different antibody clones directedagainst the same antigen molecule and even the sameMab conjugated with different fluorochromes could givedifferent ABC values by using this method (8). In addi-tion, an apparent endless avidity of QSC beads thatseems unable to saturate may introduce further variabili-ty in quantitative fluorescence measurements (9).In 1991, Poncelet et al. reported that CD4þ T cells

from HIV-infected individuals bound consistently about46,000 CD4 Mab molecules measured according to thecalibration curve generated with cell lines expressingknown amounts of CD5 molecules detected via radio-labeled CD5 Mab (5,10). Since then, CD4þ expression onT cells has been suggested and implemented as a biologi-cal calibrator for quantification of other surface antigens(11). A similar CD4 expression level on fresh normalwhole blood was also demonstrated by Davis et al. usingunimolar Leu 3a-PE conjugate and Quantibrite PE beads(7). Nonetheless, measurements of CD4 expressiondepend on variables such as fixation conditions, anti-body clones, fluorochrome and conjugation chemistries,and quantitation methods used. Because of these varia-bles, CD4 expression levels in terms of ABC valuesreported in literature vary significantly (2,6–8,12–15).Resolution of these discrepancies would be useful.In this study, we focus on CD4 ABC quantification by

employing MESF-based methods including MESF FITCmicrospheres and Quantibrite PE beads. The use ofQuantibrite PE beads and unimolar CD4-PE conjugates isa straightforward method for the determination of ABCvalues. Different sample preparation procedures areimplemented for evaluating their effects on the ABC val-ues. CYTO-TROL Control Cells, lyophilized lymphocytes,are used to determine the ABC values by the ratio of theMESF values of control cells stained with CD4-FITC Maband the MESF value of CD4-FITC Mab itself. We comparethe ABC value obtained for CYTO-TROL cells usingMESF FITC microspheres calibration method with the

443CD4 AND CYTO-TROL CELLS QUANTIFIED IN ABC USING QUANTIBRITE PE BEADS AND UNIMOLAR CD4-PE CONJUGATES

Cytometry Part B: Clinical Cytometry DOI 10.1002/cyto.b

value determined by using Quantibrite PE beads andunimolar CD4-PE conjugates. The ABC values for CD4expression on normal T lymphocytes determined byusing two MESF-based methodologies are also compared.

MATERIALS AND METHODS

Heparinized normal donor samples were obtainedfrom NIH Department of Transfusion Medicine. Thissample source is exemption approved by the institu-tional review board. Monoclonal antibodies, CD4-PEunimolar conjugate (clone Leu-3a) (Catalog No: 340586,Lot: 16533) and Quantibrite PE Quantitation kits (Cat-alog No: 340495, Lots: 50351) were purchased from BDBiosciences (San Jose, CA).1 CYTO-TROL Control Cells(Catalog No: 6604248) and CD4-FITC Mab (T4 FITC, Cat-alog No: 6603862) are products from Beckman Coulter(Fullerton, CA). A pH-sensitive fluorophore, SNARF-1,succinimidyl ester, was obtained from Molecular Probes/Invitrogen (Carlsbad, CA). A set of microspheres labeledwith different amounts of fluorescein molecules that istraceable to the NIST certified reference material, RM8640 (16), was used for quantifying fluorescein fluores-cence signals in terms of MESF values.

For labeling carrier protein-free unlabeled CD4 Mabwith SNARF-1, succinimidyl ester, Mab in PBS wentthrough buffer exchange to 0.1 M sodium carbonatebuffer, pH 8.3, by using Pro Spin columns from Prince-ton Separations. (Adelphia, NJ). Conjugation of SNARF-1to Mab was carried out according to the manufacturer-recommended procedure. After the labeling reaction,the spin column separation was performed twice toensure the separation of free fluorophores from the la-beled Mab and at the same time exchange the buffer toPBS, pH 7.2, 0.05% sodium azide.

The procedure for whole blood staining was describedpreviously (17). Briefly, the whole blood washed with13 PBS was stained with labeled antibodies for 30 minat RT. The cell suspensions were subsequently lysedwith 13 FACSTM Lysing Solution (BD Biosciences). Afterwashing twice with 13 PBS, the obtained leukocyteswere resuspended in 0.5–1 mL of PBS either with 1% fix-ative (Formaldehyde, Electron Microscopy Sciences, FortWashington, PA) or without the fixative serving as freshwhole blood samples. Mononuclear cells were obtainedfrom the whole blood (30 mL) using the Ficoll Hypaquedensity gradient separation procedure [Lymphocyte sepa-ration medium (ICN Biomedical, Aurora, OH)] (18). Asmall amount of remaining red cells were removed bylysing with ammonium chloride (ACK Lysing Solution,Biowittaker). Mononuclear cells were stained withmonoclonal antibodies for 30 min at RT. After washing

twice with PBS, the stained cells were resuspended in0.5–1 mL of PBS either with or without 1% fixative. Forobtaining enriched cells, 100 lL of CD19 microbeadsfrom Miltenyi Biotec. (Auburn, CA) was added to a 2 mLof whole blood sample and incubated for 15 min at 48Cfollowed by addition of a 15 mL of 1X PBS/0.5% BSA (w/w) and centrifugation at 450g for 10 min. The superna-tant was carefully pipetted off and cell pellet was resus-pended in 13 PBS. The cell suspension was subse-quently applied to magnetic separation with the auto-MACSTM Separator (Miltenyi Biotec, Auburn, CA). Thestaining procedure for the enriched cells via negativeselection with labeled Mab is the same as that for thewhole blood samples described earlier.Labeling CYTO-TROL Control Cells with CD4-FITC

Mab followed the manufacturer-recommended proce-dure. A 100 lL volume of reconstituted CYTO-TROLcells (from an original volume of 1 mL of cells plusreconstitution buffer) was stained with a 10 lL of CD4-FITC Mab. After incubation for 30 min at room tempera-ture, 1 mL of PBS was added, and cells were analyzedusing two different flow cytometers. For labeling CYTO-TROL cells with unimolar CD4-PE conjugates, a titrationcurve was first generated to ensure cell staining undersaturation conditions. A 30-lL aliquot of unimolar CD4-PE conjugate was used for staining a 100 lL volume ofreconstituted CYTO-TROL cells at RT for 30 min. Afterwashing with PBS, the stained cells were resuspended inPBS. For the pH titration measurements, the control cellsin PBS were spun at 600g for 5 min, and cells were thenreconstituted in 0.5 mL of one of four buffer media:borate buffer (pH 9.0), PBS (pH 7.3), (2-[N-morpholino]-ethane sulfonic acid) buffer solution (MES, pH 6.2), andsodium acetate buffer (pH 5.2).The flow cytometric measurements were carried out

using a FACSCalibur (BD BioSciences, San Jose, CA) anda custom-made research cytometer. ‘‘CellQuest’’ softwareand ‘‘QuantiCalc’’ software (BD BioSciences) were uti-lized in the BDB cytometer for data acquisition and anal-ysis, respectively. For the measurements, lymphocytepopulations were gated by using 2D side and forwardscatter plots, and median log channel numbers obtainedfrom fluorescence histograms were used for the determi-nation of MESF values or number of PE molecules boundper cell. Back gating with CD3 versus CD4 was alsoused to ensure the inclusion of the total T-lymphocytepopulation.The emission spectra and the MESF assignment of

CD4-FITC Mab were performed by using a custom-made,calibrated spectrofluorimeter described previously (19).The sample cell was a semi-micro cuvette (Starna, Type9F). The laser beam was focused in the middle of thecuvette using a lens with a focal length of 60 mm.Approximately, 0.5 mL of a solution (or suspension) wasstirred using a small magnetic bar and used for the meas-urements. One potential problem associated with thissample cell is the change of illuminating intensity due toscattering by the microspheres with respect to the capil-lary sampling method reported previously (19). A few

1Certain commercial equipment, instruments, and materials areidentified in this paper to specify adequately the experimental proce-dure. In no case does such identification imply recommendation orendorsement by the National Institute of Standards and Technology,nor does it imply that the materials or equipment are necessarily thebest available for the purpose.

444 WANG ET AL.


percent systematic reductions of measured MESF valuesare expected.

The assignment of MESF value of CD4-FITC Mab wascarried out according to the published procedures (19)and based on the equality of the measured fluorescenceyields of labeled antibody solution and fluorescein stand-ard reference solution [dilutions of the standard refer-ence material, SRM 1932, from NIST (20)]. With theassumption that the molecular absorptivity is the samefor the two solutions, the equality of fluorescence yieldsis expressed as

Nfl/fl ¼ NMab/Mab; ð1Þ

where Nfl and NMab refer to the molar concentrations ofsoluble fluorescein and labeled Mab, respectively, and /denotes the fluorescence quantum yield for the two enti-ties given in the subscript. Using the reference solutionas the unit of fluorescence yield, a MESF value can beassigned to fluorescein-labeled Mab as

MESFMab ¼ Nfl

NMab

: ð2Þ

The relationship between a MESF value and an ABCvalue per cell is given by

ABC ¼ MESFcell

MESFMab

: ð3Þ

In the case of unimolar PE- Mab conjugate, an ABC valueis calculated as a ratio of the number of PE moleculesbound per cell and the number of PE molecules perMab which is one. The fluorescence properties of thePE molecule in PE-Mab conjugate were measured to bethe same as the properties of cells stained with Mabthrough direct fluorometer comparison of cells and CD4-PE solution (7). With the known ABC value obtainedthrough the use of unimolar CD4-PE conjugate, an effec-tive F/P ratio of the CD4-FITC conjugate can then bedetermined by the ratio of the MESF value measured andthe ABC value obtained [Eq. (4)]. This allows the com-parison of the MESF value measured for CD4-FITC Mab[Eq. (2)] and the effective F/P ratio calculated [Eq. (4)].

Effective F=P RatioMab ¼ MESFcell

known ABCð4Þ

RESULTS

Figure 1 shows the number of PE molecules boundper T-cell determined for numerous samples accordingto the calibration curves generated by using QuantibritePE Quantification kits and with the use of unimolar CD4-PE conjugate. Three different sample preparation proce-dures were implemented including whole blood staining,Ficoll-Hypaque density gradient protocol producingmononuclear cells, and enriched cells via negative selec-tion using CD19 Mab labeled with magnetic beads. In

addition, the effect of sample fixation was explored.Fresh samples without fixation, in general, display higherABC values than those with fixation when using thesame sample preparation protocol. For instance, themean ABC value of 48,321 (CV, 2.6%) for the freshwhole blood samples is higher than the value of 43,432(CV, 3.6%) for the fixed whole blood samples. Under thesame fixation condition, whole blood samples givehigher values than mononuclear cells. The mean ABCvalue of 43,432 for the fixed whole blood samples ishigher than the value of 36,596 (CV, 7.1%) for the fixedmononuclear cells. The latter went through more purifi-cation steps that likely affect the nature of ligand-recep-tor binding processes. The mean ABC values for freshand fixed enriched cells are 45,480 (CV, 4.5%) and43,336 (CV, 3.9%), respectively; both are very close tothe values for the respective whole blood samples asexpected. Importantly, the number of PE moleculesbound per cell determined for fresh whole blood sam-ples, 48,321, is consistent with the value, *48,000,reported by Davis et al., with the use of the same quanti-fication method (7).CYTO-TROL Control Cells were stained with CD4-FITC

Mab following the manufacturer-recommended proce-dure. On the basis of the calibration curves generated byusing FITC-labeled microspheres traceable to NIST RM8640, the mean MESF value of stained CYTO-TROL cellswas determined to be 40,292 (CV, 4.0%) for six separateexperiments. Moreover, by implementing the calibration

FIG. 1. The number of PE molecules per cell determined on the ba-sis of calibration curves generated by using QuantiBrite PE Quantifica-tion kits and the use of unimolar CD4-PE conjugate. Three differentsample preparation procedures were implemented including wholeblood, mononuclear cells, and enriched cells by negative selectionusing magnetic CD19 Mab. Additionally, fixation effect was investi-gated through comparison of fresh and fixed sample measurements.



curve attained with a series dilution of the fluoresceinSRM 1932 (Fig. 2), we assigned a MESF value to CD4-FITC Mab. The equation shown in Figure 2 allows thedetermination of the equivalent fluorescein concentra-tion, Nfl, in Eq. (1) given in the ‘Materials and Methods’section using the measured fluorescence signal fromCD4-FITC Mab solution (FS). With a known concentra-tion of CD4 Mab provided by Beckman Coulter, theMESF value of CD4-FITC antibody was calculated to be1.16 according to Eq. (2). This MESF value could inheritas much as 10–15% uncertainties because of variation inthe concentration of CD4-FITC Mab. An accurate con-centration could in principle, be determined by aminoacid analysis (21). The ABC value for CD4-FITC Mab wasthen calculated to be 34,734 on the basis of Eq. (3)(Table 1). To make sure that CD4 receptor folding onCYTO-TROL cell surface is in a similar fashion as that inwhole blood samples, we determined the ABC value ofCYTO-TROL cells by using Quantibrite PE calibration

method and unimolar CD4 PE conjugate. The ABC valueis obtained to be 43,090 (CV, 2.6%) for five separateexperiments (Table 1).It is well known that fluorescein displays pH-sensitive

fluorescence (22,23). Therefore, we speculate that as afirst approximation, fluorescein molecules on the outersurfaces of lymphocytes experience the same pH micro-environment as fluorescein attached to the antibodies orin bulk solution. To verify this approximation, SNARF-1,a pH sensitive fluorophore (24,25), was conjugated tothe carrier protein-free unlabeled CD4 Mab in the samefashion as the FITC conjugation to the Mab. Figure 3exhibits emission spectra of SNARF-conjugated CD4 Maband whole blood stained with the labeled antibodies,both in 13 PBS, pH 7.3. On the basis of the ratios of thetwo fluorescence peaks at 582 nm and 645 nm, the pHvalues were estimated to be *8.0 for SNARF-labeledMab and 7.1 for the antibody-stained whole blood,respectively. The pH value that fluorescein molecules areexperiencing when attached to CD4 Mab is higher thanthe value of the buffer medium. On the other hand, thepH value that fluorophores are sensing on cell surfacesis slightly lower than that of the medium.To obtain estimation as to the percentage decrease of

the fluorescence signal due to the pH change at the cellinterface, we performed these titration measurementsusing identical number of CYTO-TROL cells in the fourdifferent buffer media. Figure 4 shows the titration curvewith respect to fluorescein solution. The titration curvesfor three brightly fluorescent FITC-labeled bead popula-tions are also included. The signals at highest pH value,*9.0, are normalized to be 1. All curves trail a similartrend except that the signals from stained CYTO-TROLcells drop more severely at low pH. Based on thesemeasurements, a *15% decrease of fluorescence signalis expected from single FITC-labeled CD4 Mab to CYTO-TROL cell stained with single CD4 antibody.We further determined MESF values for numerous nor-

mal blood samples stained with CD4-FITC Mab followingvarious sample preparation procedures. These values aregiven in Figure 5, showing trends similar to CD4 expres-sion levels measured by using unimolar CD4-PE conju-gate displayed in Figure 1. Fresh prepared whole bloodgives the highest mean MESF value 25,701 (CV, 2.2%).

FIG. 2. A calibration curve attained on the calibrated spectrofluorim-eter using a series dilution of the NIST fluorescein standard referencematerial, SRM 1932. The solid line is the linear fitting curve of the ex-perimental data points (solid circles) via the expressed equation thatrelates measured fluorescence signal, FS, to soluble fluorescein con-centration (C). This equation allows the determination of the equiva-lent fluorescein concentration for CD4-FITC Mab, which is pointed bythe arrow intersected with y-axis.

Table 1ABC Values Obtained for CYTO-TROL Cells and Different Sample Preparations by Using Both CD4-PE and CD4-FITC Conjugates

Samples

ABC

CD4-PE CD4-FITCCD4-FITC revised15% for pH effect

CYTO-TROL 43,090 34,734 40,864Fresh whole blood 48,321 22,156 26,066Fixed whole blood 43,432 17,879 21,034Fresh mononuclear cells 42,001 19,292 22,697Fixed mononuclear cells 36,596 ND NDFresh enriched cells 45,480 22,051 25,942Fixed enriched cells 43,336 20,187 23,749

ND means ‘‘not done.’’

446 WANG ET AL.


The mean values for fixed whole blood and fresh pre-pared mononuclear cells are 20,740 (CV, 1.6%) and22,379 (CV, 7.2%), respectively. As expected, the meanMESF value for fresh enriched cells, 25,579 (CV, 4.8%), isabout the same as that of fresh whole blood samples.However, the mean value for the fixed enriched cells,23,417 (CV, 3.6%), is slightly higher than the value forthe fixed whole blood. Note that the mean MESF valueof fresh prepared whole blood, 25,701, is lower thanthe mean value determined for stained CYTO-TROL cells(40,292). DISCUSSION

The mean number of PE molecules bound per celldetermined by using unimolar CD4 PE conjugate is48,321 for fresh prepared whole blood samples (Fig. 1).This value is about the same as the value reported byDavis et al. (48,000) who employed the same calibrationmethodology (7). In addition, this value is consistentwith the value, 46,000, published by Poncelet et al.according to the calibration curve generated with celllines expressing known amounts of CD5 moleculesdetected by radio-labeled CD5 Mab (9). The use of twodifferent QFC methods for obtaining similar ABC valuesenhances the measurement confidence in quantificationof CD4 expression. Considering that the use of unimolarCD4-PE conjugate is critical in the present method, thereliability of the nominal 1:1 molar ratio of the conjugatemust be ensured by the manufacturer. This study alsodemonstrated that fixation and additional sample prepa-ration steps clearly resulted in lower ABC values (Fig. 1).This suggests that fresh and fixed cell preparationsshould be compared when attempting antigen quantifica-tion. It also implies that if the CD4 lymphocyte is to beused as a biological control to normalize a given calibra-tion curve, a fresh cell preparation is needed. In fact,fresh preparations can serve as references and may beused to assign values to the fixed preparations. Theauthors realize that it may not always be possible to usefresh clinical samples due to concerns about the use of

FIG. 3. The normalized solution emission spectrum of SNARF-1 con-jugated unlabeled CD4 Mab and smoothed emission spectrum of awhole blood sample stained with SNARF-1 conjugated unlabeled CD4Mab; both were measured in 13 PBS, pH 7.3.

FIG. 4. The fluorescence signal relative to the signal at pH 9.0 as afunction of the medium pH value: fluorescein solution (open square),No. 5 FITC-labeled microsphere population (solid circle), No. 4 micro-sphere population (solid square), No. 3 microsphere population (solidtriangle), and CYTO-TROL cells stained with CD4-FITC Mab (diamond).The sigmoidal fitting curves are shown for fluorescein, No. 4 micro-sphere population, and stained CYTO-TROL cells. [Color figure can beviewed in the online issue, which is available at www.interscience.wiley.com.]

FIG. 5. MESF values determined according to the calibration curvesgenerated by using FITC-labeled microspheres traceable to RM 8460and with the use of CD4-FITC Mab. Three different sample preparationprocedures were implemented including whole blood, mononuclearcells, and enriched cells by negative selection using CD19 Mablabeled with magnetic beads. Fixation effect was also investigatedthrough comparison of fresh and fixed sample measurements.



hazardous materials. However, it should be possible to atleast explore the relationship between fresh and fixedvalues using normal blood bank samples.

We have used CYTO-TROL Control Cells stained withCD4-FITC Mab for quantifying CD4 expression level onT cells by means of a FITC MESF microsphere calibrationcurve. The mean MESF value determined is 40,292according to the calibration curves generated using FITCmicrobeads traceable to RM 8640. By dividing the MESFvalue of 1.16 assigned to CD4-FITC Mab, the ABC valueis calculated to be 34,734 (Table 1). This value is smallerthan the ABC value, 43,090, determined for CYTO-TROLcells stained with unimolar CD4 PE conjugate by usingQuantibrite PE calibration method.

One of the likely reasons for the lower value could berelated to the pH effect since fluorescein molecules dis-play pH-sensitive fluorescence (22,23). To verify thisspeculation, the pH-sensitive SNARF-1 fluorophore wasconjugated to carrier protein-free unlabeled CD4 Maband used to stain CYTO-TROL cells. The solution emis-sion spectra of SNARF-conjugated CD4 Mab and thewhole blood stained with the labeled antibodies (Fig. 3)show the pH difference between the two microenviron-ments surrounding the SNARF fluorophore. Their pH val-ues differ by *1 unit, i.e., *pH 8.0 for SNARF-labeledMab and pH 7.1 for the antibody-stained whole blood,respectively. To estimate the percentage drop in fluores-cence signals due to this pH change, we performed titra-tion measurements using CYTO-TROL Control Cellsstained with CD4-FITC Mab. As shown in Figure 4, thefluorescence signal of stained CYTO-TROL cellsdecreases by about 15% from pH 8.0 to pH 7.1. If theMESF value of CD4-FITC Mab were lowered by 15%, theABC value of CD4-stained CYTO-TROL cells would be40,864 by using Eq. (3) instead of 34,734 (Table 1). Thisvalue is reasonably close to 43,090 determined by usingQuantibrite PE quantification method. The results showthat the ABC values are fairly independent of the type offluorophore used for this cell preparation.

An equally important consideration is that FITC conju-gation generally reduces antibody affinity more thanunimolar PE conjugation. This is due to the higher num-ber of FITC fluorochromes per antibody molecule (*5fluorophores per antibody), which alters the bindingkinetics. Lower binding affinity of FITC-labeled antibod-ies could result in lower ABC values relative to the unim-olar CD4-PE conjugate. Additionally, it’s also likely thatfluorescein fluorescence quenching occurs because oftheir vicinity to the cell surface (26) and proximity toeach other (27). The PE molecule, on the other hand, iscomposed of 30 fluorophores (phycoerythrobilin andphycourobilin) that interact closely among themselvesand preserve a high fluorescence quantum yield (28).This fluorophore has a molecular weight of 243 kD andis about twice as big as an antibody molecule (150 kD).Therefore, fluorescence quenching due to the vicinity ofcell surface and amino acids of the antibody is much lesslikely; nevertheless, the spatial effect may exist becauseof the large size of the PE molecule. Ideally, radio-labeled

Mab would have a minimal perturbation on the affinitybinding interaction between surface antigen and Mab (5).An additional reason for the lower ABC value deter-

mined by means of the FITC microbead calibrationmethod with respect to Quantibrite PE Quantificationmethodology (Table 1) could be due to the refractionindex difference in Quantibrite PE beads and FITC MESFbeads. These two types of microbeads are composed ofdifferent polymers and both are different from bloodlymphocytes. Because of the difference in the index ofrefraction, we performed MESF assignments on one Cali-briteTM FITC bead population relative to the No. 4 FITC-labeled microsphere population (RM 8640) by means ofthe calibrated spectro-fluorimeter and flow cytometermeasurements. The polymer composition for Calibritebeads is similar to that of Quantibrite beads. The twoMESF values obtained for this Calibrite bead populationdiffer by less than 10%, suggesting that the refractionindex parameter in this case plays a minor role in thetwo measurement methodologies. It is well known thatthe fluorescence signal depends on the electric fieldintensity at the surface of the microsphere. We calcu-lated the electric field intensity at the distance of 10 nmfrom the surface of a sphere of radius equal to 7 3 10�6 musing the application ‘‘Near Field IE 2.00’’ from ValleyScientific, Inc. The intensity was integrated over the sur-face of the microsphere and compared to the integratedintensity of a sphere with the same index of refractionas the medium (this corresponds to the case of a homo-geneous material). For spheres of index of refractionn ¼ 1.45 and n ¼ 1.55, the difference in integrated in-tensity was less than 1%. This relatively small differencein integrated intensity would not yield perceptibleresults in the resulting fluorescence signal from micro-spheres made with poly(methyl methacrylate) and latex.Because of the facts discussed earlier, it seems more

reasonable to use an effective F/P ratio than the assignedMESF value of the antibody for the determination of theABC value according to Eq. (3). In principle, with aknown ABC value for a specific surface receptor that isestablished by different quantification methods, onecould determine the effective F/P ratios of labeled Mabsfrom different vendors by using Eq. (4). In the presentstudy, the effective F/P ratio for CD4-FITC Mab is calcu-lated to be 0.54 when we divide the mean MESF valueof 25,701 measured for the normal blood samples (Fig.5) by the generally agreed ABC value of 48,000 for CD4expression. This ratio value (0.54) is only half of theMESF value (1.16) determined by implementing rigor-ously MESF assignment procedure. This mathematicallyassigned MESF value of the Mab is referred as the effec-tive F/P ratio obtained by cuvette fluorimetry in the CLSIguideline document (29). By binding to a microbead forthe determination of the effective F/P ratio of CD4-FITCconjugate used, nonetheless, Vogt et al. have obtainedresults consistent with the *48,000 ABC consensusvalue (30). It is therefore, more appropriate to measurethe fluorescence of a conjugate bound to a biological cal-ibrator than the solution fluorescence of the conjugate

448 WANG ET AL.


for the determination of an effective F/P ratio of the con-jugate.

The ABC value of 43,090 obtained for unimolar CD4PE-stained CYTO-TROL cells is reasonably close to thevalue of 48,321 obtained for fresh whole blood. This sug-gests that the effects of lyophilization are negligible. How-ever, a revised ABC value of 40,864 for CD4-FITC Mab la-beled CYTO-TROL cells versus 26,066 for fresh wholeblood (Table 1) was unexpected and more difficult toexplain. In other words, why are the PE-labeled CYTO-TROL cells and fresh whole blood in reasonable agree-ment, whereas the FITC-labeled CYTO-TROL cells andfresh whole blood are so different. Although FITC conju-gation as discussed earlier undoubtedly changes antibodyaffinity, the most likely explanation for this close to 2-folddifference seems to be attributable to the surface interac-tion of the CD4-FITC Mab and CD4 antigen receptor onfresh whole blood cells. The result of this surface interac-tion is thought to cause fluorescence quenching as themajor explanation for this observation. We are aware thatthe CD4 PE and CD4 FITC clones used in this study arenot identical. However, the MESF values obtained for theBD CD4-FITC clone was even lower than the values forthe Coulter Beckman clone (data not shown) further sup-porting our observations. Studies are underway to addressthis issue of quenching, valency, effects of lipid rafts, andthe acidic glycocalyx loss on CYTO-TROL cells.

LITERATURE CITED1. Lenkei R, Mandy F, Marti GE, Vogt R, editors. Quantitative fluores-

cence cytometry: An emerging consensus (special issue). Cytometry1998;33(2).

2. Gratama JW, D’hautcourt Jean-Luc, Mandy F, Rothe G, Barnett D,Janossy G, Papa S, Schmitz G, Lenkei R. Flow cytometric quantita-tion of immunofluorescence intensity: Problems and perspectives.Cytometry 1998;33:166–178.

3. Hoffman RA. Standardization and quantitation in flow cytometry. In:Darzynkiewicz D, Robinson JP, Crissman HA, editors. Methods inCell Biology 63 (Cytometry), 3rd ed. San Diego, CA: AcademicPress; 2000. pp 300–340.

4. Schwartz A, Wang L, Early E, Gaigalas AK, Zhang Yu-Zhong, MartiGE, Vogt RF. Quantitating fluorescence intensity from fluorophore:The definition of MESF assignment. J Res Natl Inst Stand Technol2002;107:83–91.

5. Poncelet P, Carayon P. Cytofluorometric quantification of cell sur-face antigens by indirect immunofluorescence using monoclonalantibodies. J Immunol Methods 1985;85:65–74.

6. Bikoue A, George F, Poncelet P, Mutin M, Janossy G, Sampol J. Quan-titative analysis of leukocyte membrane antigen expression: Normaladult values. Cytometry 1996;26:137–147.

7. Davis KA, Abrams B, Iyer SB, Hoffman RA, Bishop JE. Determinationof CD4 antigen density on cells: Role of antibody valency, avidity,clones, and conjugation. Cytometry 1998;33:197–205.

8. Lenkei R, Andersson B. Determination of the antibody bindingcapacity of lymphocyte membrane antigens by flow cytometry in58 blood donors. J Immunol Methods 1995;183:267–277.

9. Bikoue A, Janossy G, Barnett D. Stabilised cellular immuno-fluores-cence assay: CD45 expression as a calibration standard for humanleukocytes. J Immunol Methods 2002;266:19–32.

10. Poncelet P, Poinas G, Corbeau P, Devaux C, Tubiana N, Muloko N,Tamalet C, Chermann JC, Kourilsky F, Sampol J. Surface CD4 densityremains constant on lymphocytes of HIV-infected patients in theprogression of disease. Res Immunol 1991;142:291–298.

11. Hultin L, Matud JL, Giorgi JV. Quantitation of CD38 activation anti-gen expression on CD38þ T cells in HIV-1 infection using CD4expression on CD4þ T lymphocytes as a biological calibrator.Cytometry 1998;33:123–132.

12. Lenkei R, Gratama JW, Schmitz G, D’hautcourt JL, Arekrans A,Mandy F, Marti GE. Performance of calibration standards for anti-gen quantitation with flow cytometry. Cytometry 1998;33:188–196.

13. Denny TN, Stein D, Mui T, Scolpino A, Holland B. Quantitativedetermination of surface antibody binding capacities of immunesubsets present in peripheral blood of healthy adult donors. Cytom-etry B (Clin Cytometry) 1996;26:265–274.

14. Serke S, van Lessen A, Huhn D. Quantative fluorescence flowcytometry: A comparison of the three techniques for direct andindirect immunofluorescence. Cytometry 1998;33:179–187.

15. Zenger VE, Vogt RF, Mandy F, Schwartz A, Marti GE. Quantitativeflow cytometry: Inter-laboratory variation. Cytometry 1998;33:138–145.

16. Report of Investigation of RM 8640 (microspheres with immobilizedfluorescein isothiocyanate) issued on February 19, 2004. The certifi-cate includes the MESF values of the reference beads and the uncer-tainties of the assignments.

17. Wang L, Abbasi F, Gaigalas AK, Vogt RF, Marti GE. Comparison offluorescein and Phycoerythrin conjugates for quantifying CD20expression on normal and leukemic B-cells. Cytometry Part B (ClinCytometry) 2006;70:410–415.

18. Renzi P, Ginns LC. Analysis of T-cell subsets in normal adults-comparison of whole-blood lysis technique to Ficoll-Hypaqueseparation by flow-cytometry. J Immunol Methods 1987;98:53–56.

19. Gaigalas AK, Wang L, Schwartz A, Marti GE, Vogt RF. Quantitatingfluorescence intensity from fluorophore: Assignment of MESF values.J Res Natl Inst Stand Technol 2005;110:101–114.

20. Report of Certificate of Analysis of SRM 1932 (fluorescein solution)issued on March 4, 2003. The certificate includes values of concen-tration, purity, and molar absorption coefficient.

21. Kim Sook-Kyung, Hancock DK, Wang L, Cole KD, Reddy PT. Meth-ods to characterize ricin for the development of reference materials.J Res Natl Inst Stand Technol 2006;111:313–324.

22. SjOback R, Nygren J, Kubista M. Absorption and fluorescence prop-erties of fluorescein. Spectrochim Acta Part A 1995;51:L7–L21.

23. Klonis N, Sawyer WH. Spectral properties of the prototropic formsof fluorescein in aqueous solution. J Fluoresc 1996;6:147–157.

24. Whitaker JE, Haugland RP, Prendergast FG. Spectral and photophysi-cal studies of benzo[c]xanthene dyes: Dual emission pH sensors.Anal Biochem 1991;194:330–344.

25. Gibbon BC, Kropf DL. Cytosolic pH gradients associated with tipgrowth. Science 1994;263:1419–1421.

26. Buranda T, Jones GM, Nolan JP, Keij J, Lopez GP, Sklar LA. Ligandreceptor dynamics at streptavidin-coated particle surfaces: A flowcytometric and spectrofluorimetric study. J Phys Chem B 1999;103:3399–3410.

27. Deka C, Lehnert BE, Lehnert NM, Jones GM, Sklar LA, SteinkampJA. Analysis of fluorescence lifetime and quenching of FITC-conjugated antibodies on cells by phase-sensitive flow cytometry.Cytometry 1996;25:271–279.

28. Gaigalas AK, Gallagher T, Cole KD, Singh T, Wang L. A multistatemodel for the fluorescence response of R-phycoerythrin. Photo-chem Photobiol 2006;82:635–644.

29. CLSI. Fluorescence Calibration and Quantitation Measurements ofFluorescence Intensity; Approved Guideline I/LA24-A (ISBN 1-56238-543-7). CLSI, Pennsylvania, 2004.

30. Vogt RF, Marti GE, Schwartz A. Quantitative calibration of fluoro-chrome intensity for clinical and research applications of leukocyteimmunophenotyping by flow cytometry. In: Tyrer HW, editor. Criti-cal Issues in Biotechnology and Bioengineering, Vol 1. New Jersey:Ablex Publishing; 1995.



Documents

Discrepancy in measuring CD4 expression on T-lymphocytes using fluorescein conjugates in comparison with unimolar CD4-phycoerythrin conjugates