Circulating human CD4⁺ T cells have intracellular pools of CCR5 molecules

To the editor:

In an article in this issue, Pilch-Cooper et al report an artifact they observed when they tried to detect intracellular CCR5 molecules by flow cytometry.¹  When they stained fixed and permeabilized cells with an anti-CCR5 antibody, they obtained a strong signal even in non–CCR5-expressing cells, with a significant off-target staining in the nucleus. They concluded that their procedure generated irrelevant antibody binding sites during the fixing and permeabilization resulting in a false-positive result and that T cells do not have large intracellular pools of CCR5. Whereas the first part of their statement is probably correct, the second one might not be.

With an alternative procedure of intracellular labeling we have previously published,²  we were able to detect CCR5 molecules in the peripheral blood CD4⁺ T cells from a wild-type (WT)/WT CCR5 donor but not from a delta32(Δ32)/Δ32 CCR5 donor, the epitope recognized by the anti-CCR5 monoclonal antibody (mAb) 2D7 we used being disrupted in Δ32 CCR5 molecules^3,4  (Figure 1A). When we transduced the WT CCR5 gene into the Δ32/Δ32 CD4⁺ T cells, we detected a signal in permeabilized cells (Figure 1B). To distinguish between surface and intracellular labeling, we first exposed these transduced CD4⁺ T cells to a saturating concentration of the unlabeled anti-CCR5 mAb 2D7, permeabilized the cells or not, and thereafter stained them with the same mAb conjugated with a fluorescent dye. Under these conditions, the permeabilized cells were labeled, whereas the nonpermeabilized cells were not (Figure 1C). Fluorescence microscopy confirmed that the staining was in the cytoplasm (Figure 1D). Moreover, overnight treatment with brefeldin A, which inhibits transport of proteins from endoplasmic reticulum to Golgi, leading to protein accumulation inside the endoplasmic reticulum, increased the amount of intracellular CCR5 molecules that were stained (mean fluorescence intensity [MFI], 33 ± 2 and 61 ± 1, P < .001, without and with brefeldin A, respectively; n = 3, Figure 1E). Finally, when we exposed circulating CD4⁺ T cells to the CCR5 ligand MIP-1β, we observed a decrease in the intensity of the labeling obtained with nonpermeabilized cells (MFI, 38 ± 2 and 24 ± 1, P = .005, without and with MIP-1β, respectively; n = 3, Figure 1F), contrasting with an increase in the intensity of the labeling obtained with permeabilized cells pre-exposed to nonconjugated 2D7 (MFI, 121 ± 8 and 190 ± 3, P = .008, without and with MIP-1β, respectively; n = 3, Figure 1F). These data are compatible with the ligand-induced internalization of cell surface CCR5 molecules.

Our data argue for the presence of intracellular CCR5 in peripheral blood CD4⁺ T cells. The reason we did not get the nonspecific binding reported by Pilch-Cooper et al may be because of differences in the method of permeabilization. Our duration of exposition to the detergent is shorter, and we used saponin instead of Triton X-100, which is known to affect proteins and to permeabilize nuclear membrane.⁵  Moreover, we did not fix the cells.

We have previously determined that about two-thirds of the CCR5 molecules are located in the cytoplasm of peripheral blood CD4⁺ T cells and one-third at their surface.²  In vivo, CCR5 ligands present in the plasma probably induce the permanent internalization of cell surface CCR5 molecules, causing part of their intracellular accumulation. Consequently, blocking the interaction between CCR5 and its ligands by the administration of a CCR5 antagonist to healthy volunteers results in an increase in CD4⁺ T-cell surface CCR5 density and a decrease in the intracellular pool of CCR5 molecules.⁶  Conversely, in case of overproduction of CCR5-binding chemokines by peripheral blood mononuclear cells, as in rheumatoid arthritis, the proportion of CCR5 molecules expressed at the surface of CD4⁺ T cells is decreased compared with the proportion of intracellular CCR5 molecules.²