Influence of 4- and 6-color flow cytometers and acquisition/analysis softwares on the determination of lymphocyte subsets in HIV infection

Lymphocyte immunophenotyping is a standard diagnostic procedure that provides valuable information for the diagnosis and monitoring of patients with cellular immunodeficiencies, such as HIV/AIDS (1, 2). During the past decade there have been major advancements in determination of lymphocyte subsets with the development of new fluorochromes, monoclonal antibodies, adoption of whole blood methodologies, single platform technologies for absolute lymphocyte subsets enumeration and use of two colors to multicolor flow cytometers along with the introduction of new softwares for acquisition and analysis (3-5). The constant up-gradation in instruments and softwares is always followed by their validations and inter-laboratory comparisons in clinical diagnostics. Six color flow cytometry though is more powerful in terms of providing more information in a single tube which reduces cost and time, but their performance under routine conditions for clinical diagnostics awaits optimization in various settings. At the same time the operating softwares for acquisition and analysis of data on the 6-color flow cytometers too requires further validation for routine lymphocyte immunophenotyping in research and clinical laboratory environment.

In the year 2004, Food and Drug Administration (FDA), Deptartment of Health and Human Services, USA, had approved a 6-color flow cytometer, FACSCanto™ from BD Biosciences with FACSCanto and FACSDiva™ softwares for the identification and enumeration of leucocyte subsets for in vitro diagnostics. While the FACSCanto software is designed mainly for the clinical applications and FACSDiva software is intended mainly for the research applications, with more flexibility of acquisition and analysis for the user. This instrument is substantially equivalent to the existing 4-color flow cytometer, FACSCalibur™ which uses the Cell Quest Pro™ software for the identification and enumeration of cells. Recently Lambert et al. (2006) compared the enumeration of peripheral blood lymphocytes using 6-color and 4-color flow cytometers from two different manufacturers (6). In our study we have also compared the performance of 4-color and 6-color flow cytometers from the same manufacturer and their respective softwares for the enumeration of lymphocyte subsets in the HIV infected individuals.

MATERIALS AND METHODS

RESULTS

Lymphocyte Immunophenotyping 4/6 Marker Analysis

The lymphocytes were initially selected on a side scatter/CD45 dot plot, followed by their separation into respective subsets, CD4 T-cells (CD3⁺CD4⁺), CD8 T-cells (CD3⁺CD8⁺), B-cells (CD3⁻,CD19⁺), and NK-cells (CD3⁻, CD16⁺56⁺). Acquisition was stopped as per CDC guidelines when 5,000 events were gated for lymphocyte count (7). The representative flow cytograms of percentage CD3, CD4, and CD8 obtained using three analytical softwares on FACSCalibur and FACSCanto flow cytometers are shown in Figure 1 and the mean absolute counts obtained for all the lymphocyte subsets are shown in Figure 2. The settings obtained with automatic 4-color and 7-color bead set-up on FACSCalibur and FACSCanto respectively, were employed and accepted for gating of lymphocytes and their subsets.

Internal/External Quality Controls and Stability of the Instrument Performance

Every run of the flow cytometric analysis included a set of normal and low internal controls. The performance of the three systems, FACSCalibur with Cell Quest Pro software (Calib), FACSCanto with FACSCanto software (Canto) and FACSDiva software (Diva) used in lymphocyte immunophenotyping over a period of ∼3 months was quite satisfactory (Fig. 3).

Our laboratory subscribes to the immunology quality assessment program of the National Institute of Allergy and Infectious Diseases, Division of AIDS (NIAID/DAIDS). We compared the performance of our three analytical software systems with the data of other participating laboratories in the program using 10 samples that were sent to our lab. Our data was highly comparable to the reported median values of the lymphocyte subsets (Fig. 4).

To evaluate the reproducibility of the analysis in two send outs, CD Chex internal control samples were run four times consecutively. The values for all the lymphocyte subsets were highly comparable in each of the runs on each system (CV < 0.5%).

Systems Comparison with Laboratory Reference Methodology

The scatter plot of data obtained using three analytical softwares showed a very good correlation of the lymphocyte percentages (Figs. 5 and 6). Overall the data shows significantly higher mean CD3 T-cell counts on FACSCalibur, Cell Quest Pro (1,864 ± 1,059 cells/μl), and FACSDiva (1,859 ± 1,044 cells/μl) as compared to FACSCanto (1,840 ± 1,040 cells/μl) (P < 0.05) (Table 1). The CD4 T-cell counts of the patients ranged from 3 to 2,370 cells/μl (as measured on FACSCalibur and Cell Quest Pro). The CD4 T-cell counts were also higher on FACSCalibur, Cell Quest Pro (885 ± 770 cells/μl), and FACSDiva (892 ± 773 cells/μl) versus FACSCanto (867 ± 767 cells/μl) (P < 0.05). FACSCalibur, Cell Quest Pro, and FACSDiva showed similar values except for CD8 T-lymphocytes where FACSDiva had significantly lower values (874 ± 429 versus 897 ± 453cells/μl, P < 0.05). The B-cell counts were unaffected when either of the instruments or softwares were used, while the natural killer (NK) cells (CD16+56 positive cells) showed similar trend like CD3 and CD4 counts with significant differences in the mean cell counts between FACSCalibur, Cell Quest Pro (240 ± 165 cells/μl) and FACSDiva (238 ± 163 cells/μl) versus FACSCanto with higher NK cell counts (260 ± 176 cells/μl). Owing to higher NK cell counts, the lymphosum (sum of CD3, CD19, and CD16+56 cells in percentage) on FACSCanto (95.60% ± 2.69) were also significantly higher than FACSCalibur, Cell Quest Pro (94.70% ± 3.26), and FACSDiva (94.67% ± 3.18) (P < 0.05) analysis.

Table 1. Paired Comparisons of Lymphocyte Subsets Measured Using Three Different Softwares on FACSCalibur™ and FACSCanto™, with FACSCanto™ Showing Significantly Lower Values for CD3 and CD4 and Higher NK Cells as Compared to FACSCalibur™, Cell Quest Pro™, and FACSDiva™ (P < 0.05)

Paired comparison	Paired differences
	t	Significance (2-tailed)
CD3Calibur – CD3Canto	5.002	0.000
CD3Calibur – CD3Diva	−0.445	0.658
CD3Canto – CD3Diva	−3.814	0.000
CD4Calibur – CD4Canto	2.161	0.034
CD4Calibur – CD4Diva	−1.689	0.096
CD4Canto – CD4Diva	−3.445	0.001
CD8Calibur – CD8Canto	1.059	0.294
CD8Calibur – CD8Diva	3.300	0.002
CD8Canto – CD8Diva	1.867	0.066
CD16+56Calibur – CD16+56Canto	−4.171	0.000
CD16+56Calibur – CD16+56Diva	0.899	0.372
CD16+56Canto – CD16+56Diva	5.745	0.000
CD19Calibur – CD19Canto	−1.756	0.084
CD19Calibur – CD19Diva	−1.747	0.085
CD19Canto – CD19Diva	−0.386	0.700
LymphosumCalibur – LymphosumCanto	−2.892	0.005
LymphosumCalibur – LymphosumDiva	0.095	0.925
LymphosumCanto – LymphosumDiva	3.293	0.002

• Calibur: FACSCalibur™ Cell Quest Pro™.
• Canto: FACSCanto™. Diva: FACSDiva™.

DISCUSSION

Recent availability of 6-color flow cytometry systems along with their acquisition/analysis softwares have provided a better opportunity for the laboratory personnel and researchers to evaluate multiple lymphocyte subsets in one tube. Along with that the introduction of new flourochromes, new monoclonal antibodies, user friendly softwares, and single platform strategies have made lymphocyte enumeration on flow cytometers fairly easy and convenient (8, 9). Some of the strategies like the use of CD45 for lymphocyte gating (3, 10) use of lyse no wash protocols (11) have been proven to be quite useful in routine determination of CD4 and CD8 T-cells in HIV infected patients. Though there has been a reported comparison of lymphocyte enumeration on a 4-color versus a 6-color flow cytometer from 2 different manufacturers earlier (6), our study has not only compared these systems from the same source (BD Biosciences) but also evaluated the available acquisition and analysis softwares on these cytometry systems using HIV-infected patient samples, where these systems find their most application. In this simultaneous comparison between the instruments and softwares we ensured to keep all of the variables consistent for the three analytical systems including; sample processing and acquisition, single operator and setting up of instruments using automatic compensation settings to avoid operator induced variations. The absolute counts based on a prepared fixed number of beads (e.g., Trucount™ tubes) depend much on the operator (12) therefore; we used white blood counts and percentage lymphocytes obtained from a hematology analyzer (dual platform) in determination of absolute counts and making comparisons. In this context, it should also be noted that if a single platform technology is applied in these systems, the results could be different from the dual platform settings. The initial gating strategy during analysis with all of the 3 software systems was also kept the same by using CD45/side scatter for initial lymphocyte selection and included CD8 dim cells in calculation of the total percentage of CD8 T-cells.

The analysis of the data showed a high colinearity on all the three software systems with the readings of FACSCalibur and Cell Quest Pro as reference (R² > 0.9). Out of all the markers tested the NK cells had a slightly poor correlation with FACSCalibur and Cell Quest Pro analyses. However, when the means were compared the data showed higher lymphocyte counts with FACSCalibur, Cell Quest Pro, and FACSDiva as compared to the FACSCanto in case of CD3 and CD4 T-cells while, the NK cells showed higher counts on FACSCanto. NK-cells have bright CD45 fluorescence but have slightly more side scattering properties than the majority of the lymphocytes (13). Therefore during the initial gating of lymphocyte cluster there could be some differences in picking of NK cells in a CD45/side scatter plot when different softwares are employed for the analysis. In addition, the fact that there is a decrease in NK cells during the course of HIV infection, a small change in enumerating their population does influence the overall lymphosum and percentage of total CD3 cells. Another population of cells, the NKT+ cells may be picked up differently by the three systems under study. Multi color analysis can be used to take advantage in the clinical conditions where enumeration of minor populations such as the NKT+ cells or γδ T cells and this can be simultaneously done in the same tube during the lymphocyte enumeration (14). In general the auto-gating algorithms isolate the populations of interest in FACSCanto for clinical applications; the FACSDiva has much more options for manual gating as compared to the FACSCanto (15). This could also be one of the reasons for the similarity of data between FACSDiva and FACSCalibur, Cell Quest Pro which too has manual gating features.

In conclusion, the difference in the type of flow cytometers and softwares does influence the enumeration of the lymphocytes subsets. Though the differences in the absolute counts or percentage of lymphocyte subsets may not appear clinically significant, statistical evaluation shows significant differences between FACSCalibur Cell Quest Pro and FACSDiva versus FACSCanto analytical software systems. In the context of similarity of counts obtained between the most widely used and conventional 4 color system, FACSCalibur, Cell Quest Pro, and FACSDiva, and the fact that FACSDiva has more options for manual gating, our data suggests that FACSDiva software should be preferred over the FACSCanto software for 6-color lymphocyte immunophenotyping. Besides that, the laboratories should also mention the instrument and software used for the specimen analysis in their routine flow cytometry reports.