How to cite this article: Demaret J, Saison J, Venet F, Malcus C, Poitevin-Later F, Lepape A, Ferry T, Monneret G. Assessment of a novel flow cytometry technique of one-step intracellular staining: Example of FOXP3 in clinical samples. Cytometry Part B 2012; 84B: 197–193.
By measuring multiple parameters on a single-cell basis, flow cytometry is a potent tool to dissect the phenotypes and functions of cell subsets. However, because this technique may be time-consuming, particularly for intracellular staining, it could be problematic for its use in daily routine or in large cohorts. Recently, a novel reagent has been developed to perform intracellular staining in one step. The objective of our study was thus to assess this new method in comparison with the reference technique by focusing on FOXP3 staining in clinical samples.
Peripheral blood was collected from 15 HIV-1-infected patients, 5 critically ill patients, and 5 healthy volunteers and stained using the two different methods. Different subsets of FOXP3 positive cells were investigated by flow cytometry.
When comparing results obtained with the two techniques, no statistical differences between the percentages of CD4+FOXP3+, CD4+CD25+FOXP3+, and CD4+CD25+CD127−FOXP3+ cells were observed. Besides, a strong correlation between percentages of CD4+FOXP3+CD25+CD127− lymphocytes measured with both techniques was found in patients (r: 0.843, P < 0.001, intra-class correlation coefficient: 0.820, P < 0.001). Importantly, flow cytometry stainings obtained with the one-step method were very robust with an excellent intra-assay precision, a better discriminative power and correct stability and reproducibility of the staining even after blood storage.
Flow cytometry is considered as a powerful tool for cellular analysis, either for studying cell phenotypes or various immune cell functions (1). However, several issues exist when considering flow cytometric intracellular stainings, especially in the context of a daily routine use and in clinical research protocols including large cohorts of patients for whom samples need to be processed immediately (2). In particular, this technique is time-consuming due to permeabilization and multiple washing steps and most importantly, results are usually poorly reproducible and hardly standardizable. This renders intracellular flow cytometry use poorly suitable for multicentric clinical studies.
Recently, a novel reagent has been developed to perform flow cytometry intracellular staining in one step (including permeabilization and staining). The aim of this study was thus to test the robustness of this reagent in daily routine conditions in clinical samples.
Accumulated evidence suggests the essential functions of regulatory T cells (Treg) in the pathophysiology of many clinical conditions such as autoimmune diseases (3), transplantations (4), severe septic shock (5), cancer (6), HIV infection (7), and in the development and maintenance of immunologic tolerance (8). However, very few studies have been conducted in large prospective clinical cohorts. This could be partly explained as Treg can only be precisely identified by flow cytometry using intracellular FOXP3 staining, the key transcriptional factor for Treg (9, 10). Even if other phenotypes (such as CD4+CD25+ or CD4+CD25+CD127−) are known to be good surrogate identification strategy for Treg (11), FOXP3 remains to date, the only marker to appropriately identify those cells (9, 12).
In this study, we thus assessed the quality and reproducibility of FOXP3 intracellular staining by using this novel one-step protocol in comparison with the well-accepted reference method in clinical samples.
MATERIAL AND METHODS
HIV-1-infected patients from the Hospices Civils de Lyon HIV cohort, located in Croix-Rousse Hospital and intensive care unit (ICU) patients from Lyon-Sud Hospital and Edouard Herriot Hospital (Hospices Civils de Lyon, Lyon, France) were included. Samples were collected in EDTA anticoagulant tubes. Analyses were performed on available residual blood after routine analyses were performed.
Fifteen HIV-1-infected patients [men = 10 (67%), median age = 50 years, inter-quartile range (IQR): 38–69] were included. They all received highly active anti-retroviral therapy (HAART) [median treatment duration = 7.65 years, IQR: 3.03–11.03] and were aviremic with a viral load (VL) at baseline below 50 copies/ml (Abbott RealTime HIV-1 assay and m2000 RT system platform). Their median CD4+ cell count at baseline was 573 cells/mm3 [IQR: 391–761].
Five ICU patients were included [men = 2 (40%), median age = 70 years, IQR: 29–79]. Three of them met criteria for septic shock diagnosis of the American College of Chest Physicians/Society of Critical Care Medicine (13) and two were trauma patients (Injury Severity Score: ISS > 25).
Five healthy volunteers [men = 2 (40%), median age = 34, IQR: 26–56], all members of the laboratory staff, were also included after informed consent was given.
No statistical differences were observed in clinical data for age (P = 0.317) and gender (P = 0.342) among HIV patients, ICU patients, and healthy donors.
Reference Method of Intracellular Staining
Samples were stained with Anti-Human Foxp3 Staining Set following manufacturer's recommendations (eBioscience, San Diego, CA). As already described (14), samples were first stained using surface markers: anti human phycoerythrin (PE)-labeled anti-CD127, PE-Texas Red (ECD)-labeled anti-CD4 and PE-Cyanine5 (PC5)-labeled anti-CD25 antibodies (Beckman Coulter, Hialeah, FL) and incubated for 15 min. Samples were then lysed using Versalyse lysing solution (Beckman Coulter) for 15 min. After a washing step, permeabilization was performed using Fix/Perm Buffer for 40 min. Following three more washing steps with permeabilization buffer and incubation with this buffer and normal rat serum for 15 min (blocking step), the FoxP3 intracellular staining using anti human FoxP3-FITC antibody (clone PCH101, eBioscience) was performed for 30 min. Samples were finally washed two times with permeabilization buffer. Rat IgG2a isotype control was used to evaluate non-specific staining.
One-Step Method of Intracellular Staining
The PerFix-no centrifuge assay Kit from Beckman Coulter was assessed. Staining of fresh whole blood was performed using PE-labeled anti-CD25, PE-Cyanine7 (PC7)-labeled anti-CD127, Alexa fluor 647 (AF647)-labeled anti-FoxP3, and Pacific Blue (PB)-labeled anti-CD4 (Beckman Coulter). FoxP3-antibodies (clone 259D) were purchased from BioLegend (San Diego, CA). According to the manufacturer's instructions, samples were first fixed with the fixative reagent and incubated for 15 min. Then, aliquots were simultaneously permeabilized thanks to the permeabilizing reagent and stained with fluorochrome-conjugated antibodies. After 60 min of incubation, samples were washed once with PBS and secondly with a solution containing formaldehyde.
One-Step Method Robustness
To study intra-assay precision, two sets of analyses were performed: (i) one sample stained 10 times and analyzed 10 times (ii) one sample stained once and analyzed 10 times. To assess the stability of the staining, expression of the markers was measured immediately after staining, then 2 h and 24 h later for five patients. To assess the effect of blood storage before staining, results from fresh whole blood or obtained after blood storage for 24 h at 4°C and at room temperature for three patients were compared.
Flow Cytometric Data Acquisition and Analysis
Cytometry analyses were performed on a NAVIOS flow cytometer using the NAVIOS software (Beckman Coulter). FOXP3 expression was examined in three different CD4+ subsets as followed: CD4+, CD4+CD25+, and CD4+CD25+CD127− lymphocytes.
Statistical analyses were performed using SPSS (version 17.0; SPSS, Chicago, IL) and GraphPad Prism (version 5.03, GraphPad Software, La Jolla, CA) software. Normality of the parameters was assessed using the Kolmogorov–Smirnov test. T-test, Wilcoxon test for continuous variables and exact Fisher test for categorical variables were performed. Correlations were studied using the Pearson's correlation coefficient test and the intra-class correlation coefficient (ICC). The Bland–Altman approach was also used to assess the agreement between the methods. P-values were considered significant when lower than 0.05.
As Treg, independently of FOXP3 expression, are usually described as either CD4+CD25+ or CD4+CD25+CD127− lymphocytes, FOXP3 expression was evaluated in these different subsets. Examples of plots representing co-expression of either FOXP3 and CD25 or FOXP3 and CD127 in CD4+ gated cells are shown in Figure 1.
No statistical differences between the percentages of CD4+FOXP3+, CD4+CD25+FOXP3+, and CD4+CD25+CD127−FOXP3+ lymphocytes measured with the two techniques were observed. This was true for the overall patient population as well as in each group individually (Table 1). In line, a strong correlation between percentages of CD4+FOXP3+ cells obtained with the two different techniques was identified in the global population (r: 0.763, P < 0.001, ICC: 0.759, P < 0.001). Likewise, percentages of CD4+CD25+FOXP3+ (r: 0.715, P < 0.001, ICC: 0.659, P < 0.001) and CD4+CD25+CD127−FOXP3+ lymphocytes (r: 0.843, P < 0.001, ICC: 0.820, P < 0.001) (Fig. 2) were nicely correlated between reference and one-step procedures. Moreover, Bland–Altman plots revealed good agreement between both methods regardless of the phenotype (Fig. 2).
Table 1. Percentages of Positive Cells Among CD4+ Lymphocytes and Mean of Fluorescence Intensity (MFI) of FOXP3 for Different Treg Phenotypes
One step method
Results were obtained in HIV-infected patients (HIV, n = 15), critically ill patients (ICU, n = 5), healthy volunteers (HV, n = 5) and in the whole cohort (Total, n = 25). Results are presented as median and IQR (inter quartile range) between brackets. P values were calculated using student T-test.
Percentages among CD4+ cells
FOXP3+ CD25+ CD127−
Mean of Fluorescence Intensity
MFI Difference: FOXP3+ minus FOXP3−
MFI Ratio: FOXP3+ versus FOXP3−
We then compared the cell separation/discriminative power of these two techniques for FOXP3+ lymphocytes gating. The signal to noise ratio was calculated for each technique by comparing the mean of fluorescence intensity (MFI) of FOXP3 positive versus negative cells among CD4+ lymphocytes (MFI ratios and differences). Interestingly, significantly different discriminative power was identified between these two techniques in each study group and in the whole study population (Table 1). We indeed observed that FOXP3+ cell identification after one-step intracellular staining was largely easier than with the reference method (Fig. 3).
When testing intra-assay precision, we observed that coefficients of variation (CV) of the percentage of CD4+FOPX3+ cells measured with the one-step technique ranged from 4 to 5.8%. Similar results were obtained when percentages of CD4+CD25+FOXP3+ lymphocytes were measured (CV between 3.3 and 6.4%).
When testing stability of the staining, we did not observe any statistical difference. We observed in five patients that mean percentage of CD4+FOXP3+ cells was 7.6% (standard deviation, SD: 1.7) at T0, 7.1% (SD: 1.1) at T+2 h, and 7.4% (SD: 1.4) at T+24 h (Fig. 4a). In the same patients, the mean MFI ratio between FOXP3− and FOXP3+ cells among CD4+ lymphocytes was 10.5 (SD: 2.2) at T0, 10.3 (SD: 1.5) at T+2 h, and 9.6 (SD: 2) at T+24 h (Fig. 4b). Similar results were obtained with other phenotypes of Treg (CD4+CD25+FOXP3+ and CD4+CD25+CD127−FOXP3+). These results indicated an excellent staining stability.
Finally, when blood storage before staining was tested, we did not observe any statistical difference. The mean percentage of CD4+FOXP3+ cells for three patients were respectively, 7.0% (SD: 0.8) at T0, 6.9% (SD: 1.6) after storage for 24 h at 4°C, and 7.1% (SD: 0.9) after 24 h storage at room temperature (Fig. 4c). The mean MFI ratio between FOXP3− and FOXP3+ cells among CD4+ population for these patients was found to be very stable at 10.7 (SD: 3.1) at T0, 11.7 (SD: 3.2) after 24 h storage at 4°C, and 10.6 (SD: 2.1) after 24 h storage at room temperature (Fig. 4d). We obtained the same results considering other phenotypes of Treg (CD4+CD25+FOXP3+ and CD4+CD25+CD127−FOXP3+).
The main result of our study is to show that Treg percentages obtained using reference and one-step techniques are well correlated. This was observed whatever the phenotype used for Treg identification (i.e., FOXP3+ cells among CD4+, CD4+CD25+, or CD4+CD25+CD127− lymphocytes). Another important result was that flow cytometry staining obtained by using one-step intracellular staining appeared as a robust technique with good intra-assay precision and adequate staining stability even after blood storage.
Interestingly, this novel technique presents with better advantages than reference technique.
Indeed, this one-step intracellular staining procedure offers several improvements. Intracellular FOXP3 staining with this technique takes about 90 min to complete (CD4, CD25, CD127, and FOXP3 stainings) whereas 200 min are needed for the usual reference method. This is due to the steps of permeabilization and staining (extra and intracellular) that are done at the same time in the one-step technique therefore shortening additional incubation times and wash/centrifugations cycles. Importantly, this diminished number of washing cycles did not impact the non-specific background fluorescence that could have been expected to be amplified. In contrast, the signal to noise ratio between FOXP3+ and FOXP3− cells was even improved with the one-step approach (Table 1). This enables for a better discrimination of the positive FOXP3 cells amongst CD4+ population with the one-step method. This is without the requirement of an isotype control use. In contrary, such specific cell separation appears as virtually impossible in some patients with the reference technique (Fig. 2).
Therefore, specifically regarding FOXP3 staining in clinical samples, our results open new perspectives of Treg investigation either in large prospective clinical studies or in daily routine. This could also be relevant for the study of different Treg subsets such as cells co-expressing CD45RA, CD38, or CTLA4 (15, 16). Moreover, this new technique obviously allows for the staining of other intracellular markers whatever the cells/phenotypes considered and could be of major interest for various users of flow cytometry.
Our study presents some limitations. In particular, two different fluorochromes and two different clones were used for intracellular FOXP3 staining in the compared techniques. This could partly explain the difference in discriminative power between the methods. For example, some studies have pointed out that the clone used in the reference method was associated with a higher level of non-specific staining compared with other clones (17, 18). This obviously needs to be confirmed in a subsequent study. Moreover, we only studied the correlation between the methods in three different pathologic situations. This novel method may therefore not be suitable in other clinical conditions. Thus, these preliminary results should be validated in a larger cohort of patients and in other clinical contexts. Besides, standardization of this novel technique should be elaborated on.
By comparing a one-step intracellular staining procedure with a reference protocol, we obtained a strong correlation between percentages of FOXP3+ Treg. Moreover, with a better staining quality, a shorter realization time and no isotype control requirement, this one-step procedure appears adequate for a daily routine use and in prospective clinical studies. This one-step procedure of intracellular staining therefore may represent an important improvement in the study of intracellular molecules by flow cytometry in clinical studies.
The authors would like to thank Anne Portier and Caroline Guignant for technical assistance.
This study was supported by Beckman Coulter through donations of laboratory equipments and supplies. This private company had no role in the study design, the collection or interpretation of the data. Similarly, Beckman Coulter had no role in the preparation of the manuscript or the decision to submit it for publication. Conversely, eBioscience did not supply any of the reagents for the reference method.