- Top of page
- MATERIALS AND METHODS
- LITERATURE CITED
- Supporting Information
FOXP3 is a key transcription factor expressed by regulatory T cells (Treg cells). However, differences in staining and analysis protocols have led to conflicting results. Moreover, the transient upregulation of FOXP3 that follows activation in non-Treg cells renders the interpretation of FOXP3 data more difficult in humans than in mice. Human peripheral blood mononuclear cells (PBMCs), isolated CD25− or CD25+CD4+ T cells were stained with three different anti-FOXP3 clones (PCH101, 206D, and 259D) alone or in combination, and using different permeabilization methods. FOXP3 expression was evaluated following T cell activation by several pathways. Gating based on a population that did not express FOXP3 (such as CD3−CD4− T cells) allowed for the optimal characterization of Treg cells. The 206D clone detected a lower percentage of cells than PCH101 or 259D. In contrast, 259D stained a population of activated T cells that PCH101 did not. Staining with two clones together consistently increased the proportion of FOXP3+ cells. However, it is likely that only the double positive cells are Treg cells, as they expressed the highest CD25 and lowest CD127 levels. Our results emphasize that the choice of staining protocol leads to very different results concerning the frequency of Treg cells in humans. A more consistent identification of these cells will improve the knowledge of their biology, particularly during disease processes. © 2010 International Society for Advancement of Cytometry
Regulatory T cells (Treg cells) play a major role in the homeostasis of the immune system (1–5). Treg cells are a subpopulation of CD4+ T cells that represent ∼1–10% of circulating CD4+ T cells (6). They were first characterized by the constitutive expression of the IL-2Rα chain (CD25) (1). More recently, it was shown that they usually express low levels of the IL-7Rα chain (CD127) (7, 8). The transcription factor FOXP3 (Forkhead box P3) plays a crucial role in Treg differentiation, function, and biology in both mice and humans (9–13). However, because FOXP3 is an intracellular protein, surrogate surface markers must be used to purify Treg cells. Additional markers are frequently associated with human Treg function, including cytotoxic T lymphocyte associated antigen (CTLA-4) (14), L-selectin (CD62L) (15), αE-integrin (CD103) (16, 17), and the glucocorticoid-induced tumor necrosis factor receptor (GITR) (18). However, none of these markers are selectively expressed by Treg cells, as they are also transiently upregulated in recently activated effector T (Teff) cells (18–21).
The first commercially available monoclonal antibody (mAb) for studies of FOXP3 in human Treg cells (hFOXY) was quickly supplanted by the more specific mouse mAbs 206D, 236A, and 259D (22) and the rat anti-FOXP3 PCH101 clone (eBioscience) in flow-cytometry applications. PCH101 targets the N-terminal region of the 431 amino acid (aa) FOXP3 protein, while 206D, 236A, and 259D bind an epitope within the remaining N-terminus (aa 105–235), near the zinc finger region of FOXP3 (23). All four of these antibodies recognize both full-length and alternatively spliced human FOXP3 (13, 24). Recent studies reported discrepant findings about FOXP3 expression in unstimulated and activated human T cells. These differences could come from the use of different anti-FOXP3 clones (25–29), different methods of cell permeabilization (25, 30), and/or the different gating strategies used to identify FOXP3+ cells (26, 28, 31).
For this purpose, we compared the proportion of FOXP3+ cells detected when stained with different anti-FOXP3 clones, conjugated with different fluorochromes, used alone or combined. We analyzed FOXP3 expression in both unstimulated and activated peripheral blood mononuclear cells (PBMCs), as well as in CD4+ T cells purified by magnetic bead negative selection or sorted Treg cells. We also determined the optimal gating strategy.
- Top of page
- MATERIALS AND METHODS
- LITERATURE CITED
- Supporting Information
Treg cells act as key regulators in the maintenance of immune tolerance and prevention of autoimmunity. Expression of FOXP3 has proven to be a reliable marker for Treg cells. However, staining with different anti-FOXP3 mAb clones has led to different results and whether one clone is better than the other is still debated. Herein, we performed an extensive study of FOXP3 staining using different staining conditions and gating definitions, to determine the optimal strategy.
Our findings clearly demonstrate that, in agreement with previous studies (34, 38, 39), using an isotype control to set the gate could lead to misleading results because their use either underestimated or overestimated the FOXP3+ cell population in both unstimulated and stimulated cells. Using CD3+CD4− or CD3−CD4− populations instead of an isotype control mAb or a FMO control to define the limits of the positive gate constituted the most reliable gating strategy, and that for all anti-FOXP3 clones. Our data are thus in agreement with those reported by Pillai and Karandikar (26). In these experiments, the PBMCs were purified from healthy subjects, and no difference was found when using either cell population as the FOXP3− population. It should be noted that in some circumstances, including HIV infection (40–45), there is an upregulation of FOXP3 in CD8+ T cells and, therefore, data are expected to be different when CD3+CD4− or CD3−CD4− cells are used to define the FOXP3− population. The inclusion of both CD3 and CD4 antibodies is, therefore, important in staining of human PBMCs, to permit a better characterization of these CD8+FOXP3+ cells. Although it has been shown that FMO controls allow to accurately determine positivity and set regions in samples containing multilabeled subpopulations (32, 33, 46–48), our results indicate that it can lead to overestimation of the percentage of FOXP3+ cells.
The comparison between different fluorochromes indicated that the anti-FOXP3 FITC-conjugated mAb (clone PCH101) detected low levels of FOXP3+ cells. The limited brightness of FITC compared to that of PE, AF647, or PB may explain these results. For this reason, FITC-conjugated mAbs should not be used to measure FOXP3+ cell frequency in complex cell populations such as PBMCs, although they worked well for Treg cell phenotyping after sorting. Comparing the three clones conjugated with the same fluorochrome, we found that 206D was the least sensitive clone. This finding is in accordance with the data reported by Law et al. (34). However, those authors also showed that Alexa Fluor 488- (AF488) conjugated 259D detected a higher percentage of FOXP3+ cells than PCH101 FITC-conjugated clone. Although absorption and emission of AF488 and FITC are close, these two fluorochromes are not identical, which may explain the difference between the two studies. In contrast to our findings, Grant et al. (25) found that 206D detected a higher percentage of FOXP3+ cells than PCH101 in unstimulated PBMCs. The different gating strategy used by these authors could explain these opposite results. Indeed, Grant et al. used an isotype control mAb to define FOXP3 staining within the CD3+CD4+CD25+ cells (25), whereas we used the CD3+CD4− cells as the FOXP3− population.
When different cell fixation/permeabilization methods were compared, our results showed that buffer#1 was the best buffer to investigate FOXP3 expression in conjunction with cytokine detection in stimulated PBMCs. However, it should be noted that this buffer does not permit the detection of some other intracellular proteins, such as the HIV core protein in infected cells (data not shown). These findings, which are in agreement with the results from a recent article (34), emphasize the fact that staining procedures should be carefully evaluated depending on the application. Moreover, a short PMA/ionomycin stimulation (6 h) decreased FOXP3 expression, in contrast to the other types of stimulation, and this downregulation should be taken into consideration when stimulated FOXP3+ cells are characterized in this context.
Higher percentages of FOXP3+ cells in unstimulated or stimulated samples were observed by staining with two FOXP3 clones used together than when one clone was used. However, the clone detecting a population of FOXP3+ cells that another clone did not, was not the same for stimulated or unstimulated cells. Indeed, PCH101 stained a minor FOXP3+ population not stained by 259D in unstimulated cells, whereas the inverse was true for stimulated cells. It has been reported that activation of Treg cells leads to the proteolytic cleavage of FOXP3 either at N-terminal or C-terminal sites inducing major topological changes and altering its DNA-binding properties (24). It is possible that such changes generate FOXP3 species better recognized by the 259D clone; such events may explain the differences we observed using 259D and PCH101. Another result deserves attention: cells stained by two clones expressed the highest levels of CD25 and the lowest levels of CD127, in unstimulated PBMCs. Therefore, staining with two clones may help increase the specificity of the staining, which could be important to phenotype Treg cells in disease processes that are accompanied with chronic activation of T cells. Although we could not confirm these data by performing functional assays with FOXP3+ cells, it has previously been shown that sorted human CD25+CD127low and CD25hiCD127low displayed all the characteristics of functional Treg cells such as high expression of FOXP3 and CTLA-4, the ability to suppress proliferation of other T cells, as well as hyporesponsiveness to TCR stimulation (49).
Collectively, our results indicate that the choice of gating strategy is crucial; in particular, the use of an anti-FOXP3 isotype control or FMO to gate could lead to misleading results. Moreover, depending on the activation status of the cells, the optimal detection of FOXP3+ cells was not achieved with the same clone. In summary, the use of two anti-FOXP3 clones in combination may improve both the sensitivity and the specificity of FOXP3 staining in complex cell populations.