Comparison of fluorescein and phycoerythrin conjugates for quantifying CD20 expression on normal and leukemic B-cells

Authors

  • Lili Wang,

    Corresponding author
    1. Biochemical Science Division, National Institute of Standards and Technology (NIST), Gaithersburg, Maryland 20899-8312
    • National Institute of Standards and Technology, 100 Bureau Drive, Stop 8312, Gaithersburg, MD 20899-8312, USA
    Search for more papers by this author
  • Fatima Abbasi,

    1. Center for Biologics Evaluation and Research, U.S. Food and Drug Administration, NIH Building 29B, Bethesda, Maryland 20892
    Search for more papers by this author
  • Adolfas K. Gaigalas,

    1. Biochemical Science Division, National Institute of Standards and Technology (NIST), Gaithersburg, Maryland 20899-8312
    Search for more papers by this author
  • Robert F. Vogt,

    1. Division of Laboratory Sciences, CDC, Atlanta, Georgia 30341
    Search for more papers by this author
  • Gerald E. Marti

    Corresponding author
    1. Center for Biologics Evaluation and Research, U.S. Food and Drug Administration, NIH Building 29B, Bethesda, Maryland 20892
    • Center for Biologics Evaluation and Research, U.S. Food and Drug Administration, NIH Building 29B, Room 2NN08, Bethesda, MD 20892, USA
    Search for more papers by this author

Abstract

Background

Numerous methods for quantitative fluorescence calibration (QFC) have been developed to quantify receptor expression on lymphocytes as potential disease biomarkers. CD20 expression in B-cell chronic lymphocytic leukemia (B-CLL) is one of the best examples of such a biomarker, but results from the use of different QFC methods vary considerably.

Methods

We measured CD20 expression on normal and B-CLL B-cells, using FITC and PE conjugates from the same monoclonal antibody (Mab). As a biological control and calibrator, we also measured CD4 expression on T-cells with FITC and PE Mab. Calibration curves were constructed using the CLSI (formerly NCCLS) consensus guidelines for QFC. Calibration with QuantiBRITE™ PE-labeled microspheres and the use of unimolar PE conjugates provided direct measurement of antibody bound per cell (ABC) for CD4 and CD20. Calibration for FITC conjugates was based on molecules of equivalent soluble fluorochrome (MESF), as determined by NIST RM 8640 microsphere standards. These MESF values were then converted to ABC, using the CD4 T-cell as a biologic calibrator, to normalize FITC and PE results for CD20 expression.

Results

On normal B cells, the mean ABC value for unimolar CD20-PE conjugate was 143,500 (CV ± 19.1%). The mean ABC value for B-CLL B-cells stained with the same conjugate was 21,700 (CV ± 42.0%). Using the CD4 T-cell as a biologic calibrator for FITC conjugate, the mean ABC value for CD20-FITC on normal B-cells was 199,300. CD20-FITC staining on B-CLL cells was generally too weak for accurate quantification. On normal T-cells, the mean ABC value for CD4 unimolar PE conjugate was (36,800 ± 10.4)%, and it did not differ significantly in CLL samples.

Conclusion

The expression of CD20 on normal and B-CLL lymphocytes can be quantified in ABC units using unimolar CD20-PE conjugates. In addition, CD4 expression on T-cells can be used as a biological calibrator to quantify CD20-FITC ABC, with reasonable agreement between the two conjugates with different fluorochromes. Issues regarding the accuracy of MESF microsphere calibrators and effective F/P ratios for FITC conjugates will require additional laboratory studies. © 2006 International Society for Analytical Cytology

B-cell chronic lymphocytic leukemia (B-CLL) is a neoplastic disease characterized by the accumulation of small to medium-sized, mature-appearing lymphocytes in the blood, bone marrow, and lymphoid tissues. It is the most common hematologic malignancy in adults in the western world, and accounts for 30% of all leukemias. Its incidence increases logarithmically from age 35 years and does not reach a plateau. Of all the hematologic malignancies, B-CLL shows the highest familial clustering, suggesting a genetic component to its etiology. The diagnosis of B-CLL is made on the basis of clinical presentation, laboratory studies, and examination of the blood film. Flow cytometry plays a critical role in the process. At present, flow cytometry has been used to establish an immunophenotypic profile of B lymphocytes, that is characteristic of CLL, involving coexpression of CD5, CD19, CD20, and CD23 with low to absent expression of CD79b and FMC7. Low surface Ig expression and downregulation of CD20 are consistent finding in B-CLL (1–4). However, the reported number of CD20 receptors per cell varies by more than 1 order of magnitude. Quantification of this significant disease marker (CD20) through the development of reference standards, controls, and standardized protocols would increase its clinical usefulness. One potential therapeutic implication concerns the infusional toxicity of Rituxan®, the CD20 therapeutic monoclonal antibody (Mab), which may be related to higher levels of CD20 antigen expression (5).

In the present study, unimolar CD20-PE Mabs were used for the determination of CD20 expression on B lymphocytes, using QuantiBRITE™ PE quantitation kits as the cytometric calibration beads. The CD20 quantification was carried out on both normal B cells and B-CLL patient samples. We utilized CD4 PE as a positive control and gating reagents that define major lymphocyte as subsets: T cells, B cells, and NK cells. CD4 was chosen because it has been studied extensively (6–9), shown little variation between normal subjects, and demonstrated excellent baseline separation. The validation efforts on CD4 quantification should ensure the confidence of quantifying the B-cell marker, CD20. We further applied the methodology described previously (10) to measure molecules of equivalent soluble fluorophore (MESF) to fluorescein-labeled CD4 and CD20, using a reference MESF microbead labeled with fluorescein as the calibration standard for flow cytometers. Using the antibody bound per cell (ABC) values obtained with unimolar CD4-PE and CD20-PE conjugates, we determined whether CD4 T-cells could be used as a biological calibrator for converting FITC MESF values to ABC values for the CD20-FITC conjugate.

MATERIALS AND METHODS

Heparinized normal donor samples were obtained from NIH Department of Transfusion Medicine. Chronic lymphocytic leukemia (CLL) samples were from NCI outpatient clinic. Both sample sources are approved by the institutional review board. Monoclonal antibodies (Mabs), CD4-PE unimolar conjugate (Catalog Number: 340586, Lot: 16533), CD20-PE unimolar conjugate (Catalog Number: 347201, Lot: 14549), CD4 FITC (Catalog Number: 340133, Lot: 61086), standard grade CD4 PE (Catalog Number: 347327, Lot: 61086), and standard grade CD20 FITC (Catalog Number: 347673, Lot: 58175) were purchased from BD Biosciences (San Jose, CA)1. For PE measurements, QuantiBRITE™ PE Quantitation kits (Catalog Number: 340495, Lots: 64511, 59998, and 23829) were also obtained from BD Biosciences. A NIST-certified reference material, RM 8640 (11, 12), was used for quantifying fluorescein fluorescence signals in terms of MESF values.

The procedure for whole blood staining was described previously (10). Briefly, the whole blood, either not washed or washed with 1× PBS, was stained with labeled antibodies for 30 min at 4°C. The stained cells were subsequently lysed with 1× FACS™ Lysing Solution. After washing twice with 1× PBS/0.1% sodium azide, the obtained leukocytes were resuspended in 0.5–1 mL of 1% fixative in PBS (Formaldehyde, Electron Microscopy Sciences, Fort Washington, PA). Mononuclear cells were obtained from the whole blood (30 mL), using the Ficoll-Hypaque density gradient separation procedure [Lymphocyte Separation Medium (ICN Biomedical, Aurora, OH)] (13, 14). A small amount of remaining red cells were removed by lysing with ammonium chloride (ACK Lysing Solution, Biowittaker). Mononuclear cells were stained with Mabs for 30 min at 4°C. After washing twice with PBS/0.1% sodium azide, the stained cells were resuspended in 0.5–1 mL of 1% fixative in PBS.

The flow cytometric measurements were carried out using a FACScan from BD Immunocytometry Systems (San Jose, CA) and an in-house research cytometer. “CellQuest” software and “QuantiCal” software (BD BioScience) were utilized in the BD cytometers for data acquisition and analysis, respectively. For the measurements, lymphocyte populations were gated by using 2D side and forward scatter plots, and median log channel numbers obtained from fluorescence histograms were used for the determination of MESF values or number of PE molecules bound per cell. Back gating with CD45 versus CD16 was used to define total lymphocytes, taking into account the dim CD45 expression usually seen with CLL lymphocytes. We are aware that CLL lymphocytes are almost always dimmer in their CD45 expression as are normal B cells compared to normal T cells. In three-color experiments, CD19 was used to define the B cells, for which CD20 was determined. Fixed samples from mononuclear cell preparations were stored overnight in the refrigerator and the measurements were repeated in an identical manner on the same flow cytometor the following day.

The relationship between a MESF value and an ABC (the number of antibodies bound per cell) value is given by

equation image

The effective F/P ratio of the unimolar PE conjugate is 1.0, under the assumption that the fluorescence yield of the PE conjugate is the same as the fluorescence yield of the PE QuantiBRITE microbeads. An effective F/P ratio of the CD4-FITC conjugate can then be determined by the ratio of the MESF value measured and the ABC value obtained through the use of unimolar CD4 PE conjugate. Using CD4 as a biological calibrator means that the equality of the two effective F/P ratios for both CD4 FITC and CD20 FITC conjugates is assumed, in which the two conjugates are produced by the same company via a consistent labeling protocol. Hence, the ABC value of CD20 FITC conjugate is calculated as the ratio between the MESF value measured and the effective F/P value of CD20 FITC that is the same as the value of CD4 FITC.

RESULTS

The number of PE molecules determined based on calibration curves generated using QuantiBRITE PE Quantification kits is shown in Figure 1 for standard grade CD4 PE Mab. Although three different kinds of sample preparation procedures (washed in purple, unwashed whole blood in red, and mononuclear cells in blue) were implemented, the number of PE molecules obtained shows little overall variation, with a mean value of 34,400, and a coefficient of variation (CV) of 5.5%. The technical replicates of normal individual and B-CLL patient samples (data not shown) exhibit little variation in the following day.

Figure 1.

The number of PE molecules per cell determined based on calibration curves generated using QuantiBRITE PE Quantification kits. Both standard grade and unimolar CD4 PE conjugates were used for cell staining as well as three different sample preparation procedures (washed and unwashed whole blood and mononuclear cells) were implemented.

The ultimate goal of quantitative cytometric measurements is to quantify surface antigen expression level through determination of ABC value for a specific antibody against the antigen of interest. To do so, we employed unimolar CD4-PE conjugate to quantify the CD4 expression on T cells. The underlying assumption is that, when using the unimolar conjugate, the number of PE molecules determined is equivalent to the ABC value. According to the calibration curve obtained using QuantiBRITE PE microbeads, Figure 2 shows an example for the determination of the ABC value of CD4+ T cells (top figure). The ABC values measured using the unimolar conjugates are shown in Figure 1 and are also listed in Table 1 for mononuclear T cells from healthy individuals and B-CLL patients. The technical replicates measured on the following day using the NIST research cytometer equipped with a linear amplifier for data acquisition (data not shown) are consistent, but slightly higher than the results shown and done by using a FACScan flow cytometer. The mean ABC value for CD4+ T cells is 36,800 with a CV of 10.4%. Interestingly, the ABC values, determined using CD4-PE unimolar conjugate, are close to the number of PE molecules obtained using the standard grade (“off the shelf”) CD4 PE Mab (Fig. 1). For example, two samples, N5 and N6, have about the same number of PE molecules per cell. On the other hand, samples L1 and L2 have slightly more PE molecules with the use of unimolar CD4-PE conjugate. The blue bars for N7 and L3 are the results of technical replicates measured in two consecutive days.

Figure 2.

Representative single parameter histograms are showed for unimolar CD4-PE and unimolar CD20-PE conjugates. The ABC values for normal CD4+ T cells (top figure), normal CD20+ B cells (middle figure), and CLL patient CD20+ B cells (bottom figure) were determined according to the calibration curve generated using a QuantiBRITE PE Quantitation kit. The vertical and horizontal lines with arrow shows the conversion of the channel number to the interpolated ABC values based on the calibration curve.

Table 1. MESF and ABC Values Obtained from Normal Blood Donors and B-CLL Patientsa
SamplesbMESFCD4 FITC (×1000)No. of PECD4 PE (×1000, unimolar)MESFCD20 FITC (×1000)No. Of PECD20 PE (×1000, unimolar)ABCCD20 FITC (×1000)
  • a

    MESF values deter mined based on calibration curves generated by using NIST RM 8640, the number of PE molecules per cell obtained based on calibration curves generated by using QuantiBRITE™ PE Quantification kits and unimolar PE conjugates, and calculated antibody bound per cell values for CD20 FITC using CD4 expression on T cells as the biological calibrator.

  • b

    Mononuclear cell samples were used for staining. The mean MESF values were calculated to be 15,200 for CD4 (CV, 14.8%), 81,700 for normal CD20 (CV, 13.3%), and 9,500 for CLL CD20 (CV, 26.2%). The mean number of PE molecules bound per cell (ABC) was 36,800 (CV ± 10.4%) for unimolar CD4-PE stained T cells, 143,500 (CV ± 19.1%) for unimolar CD20-PE stained normal B cells, and 21,700 (CV ± 42.0%) for unimolar CD20-PE stained B-CLL samples, respectively. ND, not done.

Normal
 Donor 117.140.385.1116.8200.6
 Donor 218.533.990.5166.6165.8
 Donor 313.538.869.5163.3199.7
 Donor 416.136.7ND170.9ND
 Donor 511.336.0ND107.0ND
 Donor 6ND33.9ND136.3ND
B-CLL
 Patient 115.536.57.919.918.6
 Patient 217.240.312.423.829.0
 Patient 316.045.18.318.123.4
 Patient 4ND36.1ND35.3ND
 Patient 5ND32.8ND25.6ND
 Patient 6ND31.2ND7.7ND

Using the same methodology for unimolar CD4-PE conjugate, we obtained ABC values for unimolar CD20-PE conjugate, also shown in Table 1, for both normal donors (middle panel in Fig. 2) and B-CLL patients (bottom figure in Fig. 2). The mean values are 143,500 for healthy individuals with a CV of 19.1% and 21,700 for B-CLL patients with a CV of 42.0%, respectively. As previously noted, the surface CD20 expression level on B cells is downregulated in B-CLL patients respective to the level in healthy blood donors (1–4).

Because of the availability of the NIST RM 8640, we attempted to make MESF assignments to CD4 FITC-stained T cells and CD20 FITC-stained B cells. Examples are shown in Figure 3 for a CD4 FITC-stained CLL patient sample (top figure) and a CD20 FITC-stained normal donor (bottom figure). A typical calibration curve obtained using RM 8640 (Fig. 3) displays a linear relationship between FL1 channel numbers and MESF values for the five-microbead populations (solid circles). Using these calibration curves, the mean channel numbers for CD4+ T lymphocytes and for CD20+ B cells were converted to MESF values (Table 1). Similar to the results carried out using unimolar CD4-PE and CD4-PE conjugates, CD4 MESF values obtained for healthy donors and B-CLL patients are close to each other, with a mean MESF value of 15,200 and a CV of 14.8%. In normal donors, the mean MESF value of CD20 FITC was 81,700 with a CV of 13.3%. Out of three pairs of normal donor and CLL patient samples, the signal from one CD20 FITC-stained CLL patient sample was very dim, yet we were able to separate a low-expressed CD20+ population from a normal population. The other two B-CLL patient samples displayed a relatively distinct downregulated CD20+ expression pattern in the histogram that allowed for the determination of the MESF values of CD20+ B cells (Table 1).

Figure 3.

Representative single parameter histograms for CD4 T cells from a B-cell CLL patient (top figure) and CD20 B cells from a normal blood donor (bottom figure). A representative calibration curve is shown for the assignment of the corresponding MESF values, using the NIST RM 8640 microbeads. The CD4 histograms for normal blood donors are very similar to that for B- cell CLL patients (top figure), except for much higher CD4+ event counts. The vertical and horizontal lines with arrow shows the conversion of the channel number to the interpolated MESF values, based on the calibration curve.

DISCUSSION

In the present study, CD4 is used as a positive control that takes advantage of the relatively low interpersonal variation in the expression of CD4+ on T lymphocytes. Validation of CD4 expression reliably and accurately would increase confidence in the CD20 measurements. It is known that CD20 is downregulated in B-CLL patients, and normal blood donors have fewer B cells than do B-CLL patients. Therefore, the use of PE-labeled Mab will enhance the measurement accuracy. The averaged number of PE molecules (Fig. 1) determined using standard grade anti-CD4 PE and based on calibration curves generated by QuantiBRITE PE Quantification kits is 34,400, with a CV of 5.5%. This value is comparable to the value reported by Davis et al., ∼38,000 PE molecules, using standard grade CD4 PE from BD Biosciences and freshly prepared blood samples without fixation step (8). The authors showed that the number of PE molecules determined depended on fixation condition, and varied with CD4-PE conjugates obtained from different sources.

The use of unimolar PE antibody conjugates and PE-labeled microbeads to calibrate ABC values depends not only on the one-to-one ratio of the conjugate but also on the equivalent fluorescence yield of PE conjugated to antibodies and PE on the microbeads. This relationship has been well documented for the reagents and microbeads used in this study, but it may not apply to all PE conjugates. The ABC values obtained using unimolar conjugates are given in Table 1. The mean value determined for unimolar CD4-PE conjugate (36,800) is slightly larger than that for the standard grade CD4 PE (34,400); both are consistent with the reported values (8). Though our values are lower than the reported “ideal” value, 50,000 (15, 16), they are largely dependent on sample preparation protocols, namely, the fixation procedure used in the present study. The ABC values are relatively constant between healthy blood donors and CLL patients.

We adopted the same methodology to measure CD20 expression levels on B cells, in terms of the ABC values shown in Table 1 for normal donors and B-CLL patients. The mean value obtained for CLL patients (21,700) is more than six times lower than that for healthy individuals (143,500). The values for healthy blood donors are consistent with those reported by Bikoue et al. (6) and by Gratama et al. (17). It is worth noting that quantitative measurements of ABC values of CD20 can only be carried out using unimolar conjugate of CD20 PE, because the commercially available standard grade CD20 PE contains a small fraction of unlabeled antibodies (personal communication, Robert Hoffman, BD Biosciences). Unlabeled antibodies have higher binding affinity than PE-conjugated antibodies, resulting in lower ABC values. Considering the use of different reagents, such as the standard grade CD20 PE from BD and different quantification methods, i.e., Quantum Simply Cellular and MESF approaches, in Refs.1–4, it is difficult to compare CD20 expression levels on normal B lymphocytes in terms of the ABC values. The present study shows much less variation in CD20 expression among numerous B-CLL patients (42.0% CV).

Since CD4 is highly expressed on T lymphocytes, we also carried out CD4 quantification using fluorescein-labeled Mab and RM 8640 as the cytometric calibration standard. Based on the calibration curves generated, MESF values were obtained for the CD4 determinations. These values shown in Table 1 are also relatively consistent between normal blood donors and B-CLL patients, with a mean MESF value of 15,200. According to the same calibration curves, the mean MESF value of CD20 FITC was calculated to be 81,700. Note that a reasonable linear relationship y (y = 1.80x, R2 = 0.87) appears between the MESF values determined for CD20 FITC antibodies and the number of PE molecules measured for unimolar CD20 PE conjugates (Fig. 4). To resolve the number of antibodies bound per cell (ABC), a MESF value of FITC-labeled antibodies (or effective F/P value) is needed which is not available commercially. Given that the sample preparation protocols are the same for PE-labeled and FITC-labeled antibodies, we calculated an effective F/P value as a ratio between the measured MESF value for the CD4 FITC-stained lymphocytes (15,200) and the ABC value for unimolar CD4 PE-stained cells (36,800), which is about 0.41. Using this value (0.41) and the mean MESF value obtained for CD20 FITC-labeled lymphocytes (81,700), an ABC value for CD20 FITC-stained B cells can be determined by the ratio of the two values, 81,700/0.41 ≈ 199,300. The estimated value is reasonably close to the ABC value determined using unimolar CD20-PE conjugate (143,500, Table 1). The calculated individual ABC values for CD20 FITC-stained B cells from normal donors and B-CLL patients using CD4 as a biological calibrator are in general agreement with the values measured using unimolar CD20-PE conjugate (Table 1). The results imply that the true effective F/P ratios for these particular CD4 FITC and CD20 FITC antibodies used in this study are similar, and CD4 can be used as the biological calibrator. Nonetheless, the accurate assignment of effective F/P values is under active study in our laboratory using the procedure employed to assign MESF values to microspheres with immobilized fluorescein (RM 8640) (12). The results obtained in the present study strongly suggest that consistent ABC values are possible with different fluorochromes and preparative methods, but the reagents, calibrators, and methods must be evaluated carefully to ensure such consistency. Since ABC values do not depend on the details of the fluorescence measurement, they provide a universal scale for conjugate binding and indirectly for receptor expression.

Figure 4.

The correlation between MESF values measured using CD20 FITC antibodies and the number of PE molecules determined for unimolar CD20 PE conjugates. The solid line is a linear fitting curve of the experimental data with a R2 of 0.870. (See Fig. 2 for number of PE molecules determination and Fig. 3 legend for the determination of MESF values).

Acknowledgements

The authors are indebted to Dr. Robert Hoffman at BD Biosciences for his critical reading and commenting on the manuscript.

  1. 1

    Certain commercial equipment, instruments, and materials are identified in this paper to specify adequately the experimental procedure. In no case does such identification imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that the materials or equipment are necessarily the best available for the purpose.

Ancillary