|Home | About | Journals | Submit | Contact Us | Français|
Fc receptor-like 3 (FCRL3) is a cell surface protein homologous to Fc receptors. The FCRL3 gene is present in humans but not in mice. We found that FCRL3 protein is expressed on 40% of human naturally occurring CD4+ regulatory T (nTreg) cells (CD4+CD25+CD127low). Sorted nTreg cells with the surface phenotype FCRL3+ and FCRL3− were both hypoproliferative to T cell receptor stimulation and both suppressive on proliferation of conventional T cells (CD4+CD25−) in vitro. They both expressed forkhead box p3 (Foxp3) protein, the intracellular regulatory T cell marker. However, in contrast to FCRL3− nTreg cells, FCRL3+ nTreg cells were not stimulated to proliferate by the addition of exogenous IL-2. In addition, Foxp3+ cells induced from conventional T cells by TGF-β treatment did not exhibit FCRL3 expression. These results suggest that the FCRL3+ subset of human nTreg cells identified in this study arise in vivo, and Foxp3 expression alone is not sufficient to induce FCRL3 expression. FCRL3 may be involved in human specific mechanisms to control the generation of nTreg cells.
Fc receptor-like (FCRL)4 3 is a member of a recently identified FCRL gene family encoding proteins homologous to classical Fc receptors (1-3). FCRL3 has also been designated as FcRH3, IRTA3, IFGP3 and SPAP2 (1). FCRL1−6 encode type I transmembrane proteins that contain 3−9 extracellular Ig domains and different numbers of ITAM/ITIM in the intracellular domains. Human FCRL1−5 genes are preferentially expressed in matured B cells at various stages in response to antigen exposure, whereas FCRL6 is highly expressed in T cells (4). Although the natural ligands of FCRL proteins are not known, their potential for generating intracellular signaling and their stage-dependent expression on B cells suggest roles as receptors in tuning of immune cell functions. Human FCRL genes are clustered on chromosome 1q21−23 near classical Fc receptors genes, but FCRL2−4 genes are not present at the corresponding gene locus in mice (2, 5). These human specific FCRLs may be involved in the regulation of human specific immune responses. Using a mAb specific to FCRL3, this protein was recently reported to be expressed on NK cells in addition to B cells (6). The reactivity of this mAb to T cells has not been extensively studied although a low level of FCRL3 mRNA is detectable by RT-PCR in T cell fractions (7, 8). In this study, we investigated FCRL3 protein expression on the surface of human T cells in peripheral blood using a newly generated anti-FCRL3 mAb, and found its expression on regulatory T (Treg) cells.
Naturally occurring CD4+ Treg (nTreg) cells, which characteristically express CD25 (α chain of high affinity IL-2R), are produced mainly in the thymus and engage in upkeep of self-tolerance (9-11). Forkhead box p3 (Foxp3) has been extensively studied as a transcription factor associated with the occurrence and function of CD4+CD25+ nTreg cells (12-14). Primarily in mouse studies, it has been shown that TGF-β treatment induces Foxp3+ cells from naïve T cells and that the Foxp3+ cells are functionally and phenotypically similar to nTreg cells (15). Although an exclusive surface marker for nTreg cells has not yet been identified, cytotoxic T-lymphocyte antigen 4 (CTLA-4) and glucocorticoid-induced TNF receptor family related protein (GITR) are suggested by their involvement in Treg cells suppressive activity (16, 17).
Despite their potential importance in suppression mechanisms, conclusive identification of CD4+ nTreg cells by these surface markers is problematic because other T cells can express these markers upon activation. For example, CD25 is commonly used as an activation marker of any type of T cell because CD25 expression is transiently and strongly up-regulated on activated T cells (18). CTLA-4 and GITR are also known to be induced in naïve T cells by activation (19). An additional complication is that the marker distribution is considerably different between mice and humans. In mice, Foxp3 expression is strictly correlated with suppressor activity, whereas human naïve peripheral T cells without Treg cell activity readily express low levels of Foxp3 upon TCR stimulation (20, 21). About 6−12% of mouse CD4+ T cells constitutively express CD25 and all are suppressive in the in vitro functional assay. However, the expression level of CD25 on human CD4+ peripheral T cells is rather variable and only 2−4% of human peripheral CD4+ cells with the highest level of CD25 (CD25hi) show Treg activity (22, 23). Recent searches for nTreg cell markers in human CD4+ cells revealed that down-regulation of CD127 (IL-7 receptor α chain) is associated with suppressor function in human CD4+ cells (24, 25). The CD4+CD25+CD127low surface phenotype found in 5−8% of the human CD4+ T cell population is also used for identification of CD4+ nTreg cells in human peripheral blood (26).
In this study, we found that 40% of human CD4+CD25+CD127low nTreg cells in peripheral blood express a human specific FCRL member, FCRL3, on the cell surface and that FCRL3 expression correlates with an in vitro hypoproliferative phenotype of cells in response to exogenous IL-2.
The cDNAs of human FCRL3 and other FCRLs were obtained by RT-PCR as described previously (27, 28). The cDNAs were confirmed to encode the same proteins as the following reference sequences: Genbank (http://www.ncbi.nlm.nih.gov/Genbank/) accession number, AF459634 (FCRL1), AF459633 (FCRL2), AF459027 (FCRL3), AF343659 (FCRL4), AF343664 (FCRL5), and AY513661 (FCRL6). Expression of full-length FCRLs in transfected 293T cells and expression of their extracellular domains as human IgG Fc-fusion proteins were carried out as described previously (27, 28). Production and characterization of anti-FCRL3 mAb were performed according to a series of our established protocols for obtaining mAbs to membrane proteins in their native conformation (27-30). The FCRL3-specific mAbs were screened in an ELISA using FCRL3-Fc protein as the coated antigen. Antibody reactivity was confirmed in a flow cytometry using FCRL3-transfected 293T cells.
The anti-FCRL3, H5, produced in this study, was conjugated to PE (custom labeled by Invitrogen/Molecular probe). MAbs against FCRL1 (E3), FCRL2 (B24), FCRL4 (A1), and FCRL5 (F56) were prepared by us (27, 28). The following antibodies were purchased: biotin-labeled antibodies, anti-ICOS (ISA-3, eBioscience 13−9948); FITC-labeled antibodies, anti-Foxp3 (PCH101, eBioscience 11−4776), anti-CD16 (3G8, BD 555406), anti-CD19 (HIB19, eBioscience 11−0199), anti-CD25 (BC96, eBioscience 21−0259), anti-CD38 (HIT2, BD 555459), anti-CD45RA (HI100, eBioscience 11−0458), anti-CD56 (MEM188, eBioscience 11−0569), anti-CD69 (FN50, eBioscience 11−0699), anti-CD103 (B-Ly7, eBioscience 11−1038); PE-labeled antibodies, IgG2b-isotype control (27-35, BD 555743); peridinin chlorophyll-alpha protein (PerCP)-Cy5.5-labeled antibodies, anti-CD3 (SP34−2, BD 552852), anti-CD4 (SK3, BD 341654), anti-CD5 (L17F12, BD 341089), anti-CD8, (SK1, BD 341051), anti-CD19 (SJ25C1, BD 340951), anti-CD20 (L27, BD 340955); PECy7-labeled antibodies, anti-CD4 (RPA-T4, eBioscience 25−0049), anti-CD25 (BC96, eBioscience 25−0259), anti-CD62L (DREG-56, eBioscience 25−0629); allophycocyanin (APC)-labeled antibodies, Streptavidin (eBioscience, 17−4317), anti-AITR (eBioAITR, eBioscience 17−5875), anti-CD19 (HIB19, BD 555415), anti-CD25 (BC96, eBioscience, 17−0259), anti-CD28 (CD28.2, eBioscience, 17−0289), anti-CD31 (VM59, eBioscience 17−0319), anti-CD38 (HIT2, eBioscience, 17−0389), anti-CD45RA (HI100, eBioscience 17−0458), anti-CD45RO (UCHL1, eBioscience 17−0457), anti-CD56 (B159, BD 555518), anti-CD69 (FN50, eBioscience 17−0699), anti-CD127 (eBioRDR5, eBioscience 11−1278), anti-HLA-DR (LN3, eBioscience 17−9956), anti-CTLA-4(CD152) (BN13, BD, 555855); APC-Cy7-labeled antibodies, anti-CD3 antibody (SK7, BD 557832), anti-CD62L (SREG56, eBioscience 10−0629); Pacific blue (PB)-labeled antibodies, anti-Foxp3 (PCH101, eBioscience 57−4776), anti-CD8 (OKT8, eBioscience 57−0086), anti-CD19 (HIB19, eBioscience 57−0199), anti-CD127 (eBioRDR5, eBioscience 57−1278); AmCyan-labeled anti-CD4 (SK3, BD 339187).
Buffy coats from healthy donors were obtained from the Department of Transfusion Medicine, National Institutes of Health (Bethesda, MD) or The Sanford USD Medical Center Blood Bank, according to protocols approved by the Institutional Review Boards. A total of 52 subjects from different donors were used (age range = 19−73 years, mean age = 46, male/female = 18/28). PBMCs were isolated by density gradient centrifugation over Ficoll/Hypaque (GE Bioscience). The PBMCs were freshly used or kept in culture up to 40 h in IMDM (Invitrogen) supplemented with 10% FBS (HyClone) at 106 cells/ml before the flow cytometry analysis or cell sorting. In some T cell activation experiments, the freshly isolated PBMCs (5 × 105 cells) were seeded in a 24-well culture plate (1 ml/well) and stimulated by adding 2 × 106 Dynabeads goat anti-mouse-IgG (Invitrogen 110−33) precoated with anti-CD3 antibody (clone UCHT1, BD Biosciences, 2 μg/107 beads) and anti-CD28 antibody (clone CD28.2, BD Biosciences, 2 μg/107 beads) as described previously (31).
In a typical experiment, 106 cells were incubated in 200 μl of the antibody cocktail in FACS staining buffer (PBS containing 5% FBS and 0.1% sodium azide) for 1 h at 4°C. To stain intracellular Foxp3 or CTLA-4 proteins, cells post-stained by other surface markers were fixed and permeabilized by Fixation/Permeabilization solution (eBioscience), and then stained with the mAbs against Foxp3 or CTLA-4. Flow cytometry data was acquired using FACSCalibur, LSR II Flow Cytometer (BD Biosciences) or C6 Flow Cytometer (Accuri Cytometer). At least 50,000 events were collected for three or less colors and 300,000 events for more than 4 color analysis. All the data was further analyzed by FlowJo software (Tree Star Inc.). The gating strategies included lymphocyte gating (32), CD3-side scatter gate (33), and CD4-side scatter gate (34) as described in figure legends. To determine positive and negative events for each marker, all reagents except for the one of interest (fluorescence minus one, or “FMO” controls) were used to identify expressing cells in the fully stained sample (35). Unless indicated otherwise, a positive threshold for each marker was determined so that fewer than 0.2% of cells in the counter FMO control were included in the positive ranges. Cell sorting experiments were performed on a FACSAria or a FACSVantage Cell Sorter (BD Biosciences). In some experiments, the mRNA expression of FCRL3 and β-actin in sorted cells was examined by the RT-PCR.
The mRNA expression of FCRL3 in sorted cells was examined by RT-PCR analysis according to the protocol described previously (28). For PCR, a FCRL3 primer set (5’-CAGCACGTGGATTCGAGTCAC-3' and 5'-CAGATCTGGGAATAAATCGGGTTG-3') or a β-actin primer set (5'-AGCCCTTCAGACTCGGACTC-3', and 5'-TGGGGCAGCCTAAATCTT-3') as the control was used.
To measure the suppressor activity of the sorted population of cells, the isolated cells (candidates of Treg cells) were mixed with responder CD4+CD25−CD127hi cells (effector T cells, Teff cells) that had also been sorted from the autologous PBMCs. The Teff cells were labeled with 2μM CFSE (Invitrogen/Molecular Probes) in IMDM supplemented with 1% FBS by a 10-min incubation at room temperature. Suppression assays were performed by co-culturing 4,000 CFSE-labeled Teff cells with different numbers of sorted cells (typically 4,000, 2000, 1000 and 250) in 200 μl of IMDM containing 5% human AB serum (Bioreclamation Inc.) in round-bottom 96-well microtiter plates. The cells were stimulated by adding 25,000 anti-CD3/CD28 beads per well. The bead to cell ratio for optimal cell activation was predetermined in titration experiments. After 80-h or 96-h incubation, 100 μl of supernatant was removed from each well for cytokine assays. The cells were harvested in 0.3 ml of FACS buffer containing 7-amino-actinomycin D (0.5 μg/ml, BD Biosciences) for viability staining and 20,000 FITC-conjugated beads (Bangs Laboratories Inc., #512) were used as the standard for cell counting. CFSE dilution by Teff cell division was analyzed in live cell population using an LSR II flow cytometer and FlowJo software.
The cells isolated by sorting were labeled with CFSE. The labeled cells (5,000) were incubated with or without different numbers of anti-CD3/CD28 beads (25,000, 50,000, and 100,000) in 200 μl of IMDM containing 5% human AB serum in round-bottom 96-well microtiter plates. In some cultures, exogenous recombinant human IL-2 (Biosource/Invitrogen) was added at 10 ng/ml.
Sorted CD4+CD25−CD127hi cells (5 × 105) were cultured in 1 ml of IMDM containing 5% human AB serum in 24-well culture plates for 4 days with or without 10 ng/ml of recombinant human TGF-β (eBioscience) or anti-CD3/CD28 beads (106/ml) or IL-2 (10 ng/ml).
Cytokine production in culture supernatants of the suppressor and proliferation assays were measured using a human Th1/Th2 cytokine cytometric bead array kit (BD Biosciences) or human TGF-β1 Quantikine ELISA kit (R&D Systems) according to the manufacturers’ instructions. The human cytokines measured were IL-2, IL-4, IL-5, IL-10, TNF-α (by the beads array kit), and TGF-β1 (by ELISA). IL-2 was not measured in the samples with exogenous IL-2.
We produced an anti-FCRL3 IgG2b mAb, H5, by use of a DNA immunization protocol established previously (29). We carefully examined undesired cross reactivity of H5 mAb with other FCRL proteins because there are considerable homologies between the six extracellular Ig domains of FCRL3 and the same subtypes of Ig domains of other FCRL family members (2, 6, 27, 28). H5 mAb reacted with FCRL3-Fc fusion protein in ELISA but did not react with FCRL1, 2, 4, 5, 6-Fc fusion proteins in the same assay (Fig. 1A). H5 mAb bound to the surface of FCRL3-transfected 293T cells in a flow cytometry assay but did not bind to FCRL1, 2, 4, 5, 6-transfected cells (Fig. 1B). We concluded that H5 mAb is specific to FCRL3 and not cross reactive with other FCRL proteins. H5 mAb also bound to endogenous FCRL3 protein on several human B cell lines in which FCRL3 mRNA was expressed (Fig. 1C). H5 mAb did not cross react with FcγR1 (CD64), FcγRIIb (CD32) or FcRIIIa (CD16) in single or double color flow cytometry assays using appropriate cell lines or human PBMCs (unpublished data). H5 mAb did not bind to recombinant FCRL3 proteins in Western blotting (unpublished data), indicating that H5 mAb recognizes a conformational epitope whose structure is denatured by SDS treatment. These results indicated that H5 mAb is a highly specific antibody that recognizes an epitope of the extracellular domain of native FCRL3 on the cell surface and is an ideal probe for characterizing this protein in immune cells.
We stained PBMCs from normal donors with antibodies to surface markers for B cells, T cells or NK cells together with the anti-FCRL3 mAb conjugated to PE, and then analyzed the cell population by flow cytometry (Fig. 2A). As expected from previous reports (2, 6, 8), FCRL3 protein was expressed on B cells (CD19+) and NK cells (CD56+). In addition, we detected a small subset of T cells (CD3+) expressing FCRL3 protein. B and NK cells showed only modest levels of FCRL3 signal, which is consistent with a previous report (6). In contrast, there were two distinctive peaks in the T cell fraction, indicating the presence of FCRL3+ and FCRL3− cell subsets in the peripheral T cells. Fig. 2B shows two-color staining of human PBMCs with the anti-FCRL3 and anti-CD3 (pan T), anti-CD4 (helper T) or anti-CD8 (cytotoxic T) antibodies. FCRL3+ subsets were observed both in CD4+ and CD8+ cell populations, suggesting that FCRL3 expression is not linked to the expression of CD4 and CD8 proteins. As summarized in Fig. 2C, 1.5−14.3% of CD3+ cells, 0.5−12.9% of CD4+ cells, and 2.6−24.7% of CD8+ cells expressed FCRL3. There is no association between FCRL3 and CD56 expressions on CD3+ cells (unpublished data), suggesting that the CD3+CD56+ NKT cells do not preferentially express FCRL3. To verify FCRL3 mRNA expression in FCRL3+ T cells stained by H5 mAb, we separated PBMCs into 4 fractions (P1-P4) by cell sorting based on their differential staining for cell surface FCRL3 and CD3 (Fig. 2D). RT-PCR analysis demonstrated the restricted expression of FCRL3 mRNA in the FCRL3+ cells (P2 and P4). Cloning and sequencing of the amplified DNA fragments confirmed FCRL3 sequence.
We examined FCRL3 expression kinetics on T cells in the course of activation via TCR and co-receptor using anti-CD3/CD28-coated beads (31). As shown in Fig. 3, stimulation induced transient expression of an early phase activation marker, CD69, followed by the expression of a late activation marker, CD25 in both CD3+CD4+ and CD3+CD8+ T cells. However, the frequencies of FCRL3+ cells in any cell fraction were consistently low level (<10%) during the time course, indicating that FCRL3 expression was not up-regulated or down-regulated by T cell activation. CD4+, CD8+ and CD45RO+ cell ratios in CD3+ T cells were not significantly changed by T cell activation, which is consistent with previous reports (18). Stimulation with mitogens such as PHA and Con A did not increase the level or frequency of FCRL3 expression (unpublished data).
We next asked if there is FCRL3 expression on nTreg cells. The ideal marker set for identification of nTreg cells is still under development, especially for humans. We therefore tested three widely used strategies to identify human nTreg cells in CD4+ cells: (1) intracellular Foxp3 expression (Foxp3+), (2) CD25-positive (dim to high) combined with low level CD127 expression (CD25+CD127low), and (3) the top 2% of the highest level of CD25 expression (CD25hi). As shown in Fig. 4A, nTreg cells were identified as Foxp3+, CD25+CD127low and CD25hi cells in P2, P4 and P7 gates, respectively (24-26). As the counter populations for comparison, non-Treg cells were identified as Foxp3−, CD25−CD127hi and CD25− cells in P1, P3 and P5 gates, respectively. The P6 gate defines CD25dim cells that partly contain CD25+CD127low nTreg cells. The frequency of each population in CD4+ cells is summarized in Fig. 4B.
As shown in Fig. 4, C and D, all three Treg populations (P2, P4, and P7) contained 2 distinctive subpopulations that correspond to FCRL3+ and FCRL3− cells. In contrast, the three non-Treg populations (P1, P3, and P5) were almost all FCRL3− cells. Thus, FCRL3 is preferentially expressed on nTreg cells in human CD4+ fraction. The incidences of FCRL3+ cells in CD4+ nTreg cells are 20−50% (Fig. 4D). Fig. 5 displays different types of views of the flow cytometry analysis that confirm the relationships among expressions of FCRL3 and other markers in CD4+ cells. Fig.5B shows that Foxp3 expression level is within the positive range for both FCRL3+ and FCRL3-populations in the two nTreg gates (CD25+CD127low or CD25hi), although these two cell populations show slightly different distributions of Foxp3 expression levels.
We also examined the association of FCRL3 expression with other markers including those which had been reported to subdivide CD4+CD25hi cells into two subpopulations (26, 36). As shown in Fig. 6, our results indicated that the FCRL3+ and FCRL3− nTreg cell populations are not definable by GITR, CD38, ICOS, CD31, CD62L, CD45RA, CD45RO, CTLA-4, HLA-DR, and CD103.
We next examined the regulatory (suppressor) activity of FCRL3+ and FCRL3− subsets of CD25+CD127low nTreg cells on the proliferation of non-Treg cells. According to the gating strategy shown in Fig. 7A, CD25−CD127hi (P1) and CD25+CD127low cells (P2) were isolated from the CD4+ subset of PBMCs by cell sorting. The CD25−CD127hi (P1) cells were used as responder cells (T effector cells, Teff cells) in a proliferation assay to test candidate Treg cells isolated from autologous PBMCs. We also isolated FCRL3− (P3) and FCRL3+ (P4) cells from the CD25+CD127low cell population using a second gating (Fig. 7A). The post sorting panels confirmed that these sorted cell populations fell into the same gates used for their sorting. The isolated FCRL3-(P3) and FCRL3+ (P4) cells cannot be distinguished in their CD25 vs. CD127 profiles.
As shown in Fig. 7B, addition of CD25+CD127low cells (P2), FCRL3−CD25+CD127low cells (P3) or FCRL3+CD25+CD127low cells (P4) remarkably suppressed the proliferation of the labeled Teff cells (P1), although addition of Teff cells (P1) did not block the proliferation of labeled cells. A time course experiment (Fig. 7C) and an experiment using various addback ratios (Fig. 7D) indicated that the tested FCRL3− (P3) and FCRL3+ (P4) subpopulations of CD25+CD127low cells manifested similar levels of suppressive activity as the total CD25+CD127low cells (P2). When CD25hi gating was used to identify nTreg cells, their FCRL3 expression was also independent of suppressor activity (unpublished data). CD25+CD127low cells suppressed IL-2, IL-4, IL-5 and TNF-α production by the Teff cells in the co-cultures regardless of FCRL3 expression (supplemental Table S1)5. Inhibitory cytokines, IL-10 and TGF-β, in the growth-suppressed cultures were detected only at the same trace levels as the no addback cell cultures (supplemental Table S1)5, which confirmed that CD25−CD127hi cells do not contain a significant number of adaptive Treg cells such as Tr1 or Th3 cells secreting IL-10 or TGF-β (37, 38).
In addition to the suppressor activity, a defining characteristic of nTreg cells in vitro is their relative inability to proliferate and to produce cytokines in response to TCR stimulation. Treg cells require exogenous IL-2 in addition to TCR stimulation for proliferation. Therefore, we next examined proliferation of the sorted FCRL3+ and FCRL3− cells with or without anti-CD3/CD28 beads in the presence or absence of exogenous IL-2 in vitro. Fig. 8A shows the cell divisions after 80 h in culture, monitored by the dilution of intracellular CFSE. As expected, FCRL3− (P3), FCRL3+ (P4) or the total CD25+CD127low cells (P2) showed no proliferation by anti-CD3/CD28 beads stimulation, whereas the non-Treg CD25−CD127hi cells (P1) proliferated. An increase in the ratio of stimulating beads to cells did not induce proliferation of any of the three nTreg populations, whereas CD25-CD127hi Teff cells showed similar high levels of cell proliferation at all bead/cell ratios. Fig. 8B shows the effects of exogenous IL-2 on proliferation. The anti-CD3/CD28 beads alone or IL-2 alone showed no enhancement of proliferation in the cell cultures. In sharp contrast, when cells were stimulated both with anti-CD3/CD28 beads and IL-2, FCRL3− CD25+CD127low (P3) cells showed marked proliferation activity. However, FCRL3+CD25+CD127low (P4) cells remained hyposensitive to stimulation. The CD25+CD127low (P2) cells showed moderate proliferation that approximately corresponds to the ratio of FCRL3+ and FCRL3− cells within the CD25+CD127low population (Fig. 4D). Fig. 8C shows the growth curves of different cell populations in the presence of both anti-CD3/CD28 beads and IL-2. Non-parallel growth of FCRL3+ and FCRL3− nTreg cells suggested that the difference of susceptibility to IL-2 between these two populations was maintained during the 5-day culture period. The growth of FCRL3− and the non-growth FCRL3+ nTreg cells were recognized in the majority of the sorted cell population in the scattering profiles showing larger forward-scatter and side-scatter values for growing cells (supplemental Fig. S1)5. In addition to the difference in hypoproliferative state between FCRL3+ and FCRL3− nTreg cells, IL-10 production was induced only from FCRL3− nTreg cells and not from FCRL3+ nTreg cells by exogenous IL-2 (Fig. 8D). This illustrates an additional difference between FCRL3− and FCRL3+ nTreg cells in terms of their response to exogenous IL-2. Fig. 8D also showed that neither FCRL3+ nor FCRL3-nTreg cells produced significant amounts of IL-2, IL-4, IL-5, TNF-α, or TGF-β upon anti-CD3/CD28 stimulation and further supplementation with exogenous IL-2.
We next examined FCRL3 expression on TGF-β-induced Foxp3+ cells. TCR stimulation, in addition to TGF-β treatment, is required for induction of Foxp3 in conventional T cells. IL-2 produced from the stimulated cells was also reported to contribute to the generation of Foxp3+ cells (15, 39). For these reasons, we stimulated non-Treg CD25−CD127hi cells (P1 in Fig. 7A) with anti-CD3/CD28 beads, TGF-β, IL-2 and various combinations of these factors for 4 days. As shown in the first row of panels in Fig. 9, without any stimulation, CD25−CD127hi cells retained the same phenotype throughout the incubation period. Neither Foxp3, CD25, nor CTLA-4 was induced. In contrast, as shown in the second row of panels, anti-CD3/CD28 beads activated CD25−CD127hi cells to proliferate and become granular cells with higher side scatter and markedly increased levels of CD25 and CTLA-4 expression. As expected, Foxp3 was only slightly induced by the activation. TGF-β treatment alone did not change the phenotype of CD25−CD127hi cells compared to the untreated cells. However, a combination of TGF-β with anti-CD3/CD28 beads converted a significant fraction of the stimulated cells to Foxp3-positive status. The induced Foxp3+ cells showed a marker profile similar to nTreg cells from the PBMC population, such as high levels of CD25 and CTLA-4 and lower levels of CD127. The induced Foxp3+ cells, however, did not express FCRL3. Addition of exogenous IL-2 showed no effects on FCRL3 expression. We conclude that FCRL3 expression is not induced in Foxp3+ cells produced from non-Treg CD4+ T cells in these conditions.
In this study, we found that FCRL3 protein is expressed on 20−50% of naturally occurring regulatory T cells in human peripheral blood. FCRL3+ and FCRL3− nTreg cells are both suppressive on proliferation of conventional T cells (CD4+CD25−) in vitro. However, in contrast to FCRL3-nTreg cells, FCRL3+ nTreg cells are not stimulated to proliferate by the addition of exogenous IL-2. Interestingly, the FCRL3 gene is present in humans but not in mice (5).
Dynamics and maintenance on the FCRL3 expression on nTreg cells in vivo and the mechanism leading to the generation of two subpopulations of nTreg cells remains undefined. We have not detected other markers specific to FCRL3− nTreg cells or FCRL3+ nTreg cells. Nor have we found any in vitro culture conditions that change the level of FCRL3 expression on Foxp3+ cells in a survey using conditioned media and cytokines (including 4−5 days culture in the presence of IL-2, IL-4, IL-7, IL-15, and their combinations, unpublished data). In long term cultures of T cells in the presence of IL-2, the expression of FCRL3 was gradually diluted out by an over-growth of FCRL3-negative cells after a few weeks of culture (unpublished data). We did not detect FCRL3 expression in 3 different tumor infiltrated lymphocyte cultures (CD8+, obtained from Dr. Rosenberg, unpublished data). We succeeded to induce growth of FCRL3− nTreg cells by the combined stimulation of anti-CD3/CD28 and IL-2 (Fig. 8). However, we did not detect significant FCRL3 expression in the expanded cultures from FCRL3− Treg cells (unpublished data). In addition, as shown in Fig. 9, Foxp3+ cells induced from CD4+FoxP3− cells by TGF-β treatment did not express FCRL3. Combined these data suggest that unidentified factors are necessary to induce FCRL3 expression in vivo. FCRL3− nTreg cells and FCRL3+ nTreg cells are both suppressive (Fig. 7) and show similar positive levels of Foxp3 expression (Fig. 5). Foxp3 plays a major role in the occurrence and function of Treg cells by activating or repressing many genes by forming transcription complexes with other factors (12-14). The presence of FCRL3− and FCRL3+ subpopulations of Foxp3+ nTreg cells suggest that FCRL3 expression is not solely controlled by Foxp3.
Hyporesponsiveness (anergy) to TCR and co-receptor stimulation in cell culture is an established property of nTreg cells, which can be overcome by addition of exogenous IL-2 (40, 41). Although there is a remarkable difference between in vitro and in vivo conditions for Treg cell expansion, IL-2 also plays an important role for Treg biology in vivo (42, 43). Our data revealed that FCRL3− nTreg cells but not FCRL3+ nTreg cells undergo IL-2-dependent proliferation in vitro. This suggests that nTreg expansion and maintenance by IL-2 may be associated with the FCRL3 expression in vivo in humans. Mouse nTreg cells propagated by IL-2 in vitro temporarily lose their suppressive activity but spontaneously regain the activity once IL-2 is removed (41). Such experiments using the expanded FCRL3− Treg cells (by anti-CD3/CD28 beads plus IL-2 treatment) will be needed in future studies. The association of non-responsiveness to TCR stimulation with FCRL3 expression was also observed in the CD8+ cell fraction, whereas we didn't detect suppressor activities of FCRL3+CD8+ or FCRL3−CD8+ cells in our assay (unpublished data).
An important question is why FCRL3− nTreg cells and FCRL3+ nTreg cells respond differently to IL-2 despite the expression of CD25 receptor on both populations. We detected a difference in IL-10 production after IL-2 stimulation, indicating that IL-2 likely induces a different sequence of events in FCRL3− nTreg cells than in FCRL3+ nTreg cells, leading to the dissimilarity of the response. In preliminary assays, we did not detect different levels of CD122 (IL-2Rb) or the phosphorylated form of STAT5 (by detection using anti-phospho Ab) after IL-2 stimulation between FCRL3+ and FCRL3− Treg cell populations (unpublished data). Analysis of intracellular signaling and profiling gene expression patterns induced by IL-2 will be needed to determine the differences between FCRL3− nTreg cells and FCRL3+ nTreg cells. We plan to perform a comprehensive assay using a microarray platform.
FCRL3 resembles other Ig superfamily receptors in structure, including intracellular signaling through ITAM/ITIM and tyrosine phosphorylation. These phosphorylated tyrosines are predicted to serve as docking sites for SH2 domain-containing molecules to initiate downstream signaling pathways (2). Such FCRL3-mediated intracellular signals potentially play important biological roles under the epigenetic conditions of FCRL3+ Treg cells. In preliminary experiments, we tested the effects of anti-FCRL3 mAb in PBMCs cultured in the absence/presence of secondary antibody. We also tested the anti-FCRL3 mAb coated on anti-IgG beads. Under these limited conditions, we were not able to detect significant changes in cell growth or marker phenotype of FCRL3+ nTreg cells (unpublished data). We plan to further examine phosphorylation of tyrosine residues in the signaling motifs that may be induced by engagement of FCRL3 by the anti-FCRL3 mAb.
Ex vivo generation of Foxp3+ cells from naïve T cells has been actively studied because of the potential for exploiting Treg cells for clinical use, such as reestablishing self-tolerance in autoimmune diseases. (39, 44-47). Extensive studies in mice have demonstrated that induction of Treg cells from naïve T cells using TGF-β can lead to expression of suppressive activity accompanied by Foxp3 expression. However, the accumulated evidence suggests that human Foxp3+ cells generated in response to TGF-β do not always exhibit suppressive activity (15, 46). Our finding that TGF-β-induced Foxp3+ cells do not express FCRL3 revealed a difference in marker phenotype between TGF-β-induced Foxp3+ cells and nTreg cells generated in vivo in addition to suppressive activity. Although the association of FCRL3 expression with suppressive function was not observed, seeking ex vivo conditions that generate FCRL3+ cells may allow induction of human suppressive T cells that are closer in nature to nTreg cells. A broad repertoire of self and non-self antigens can be recognized by nTreg cells in the periphery. However, specific association of a limited number of the TLR family members with nTreg cells was also reported (48). Examination of the TLR repertoire expressed on FCRL3+ and FCRL3− nTreg cells may allow elucidation of the roles of these cells in regulating the immune system. Other important populations of Treg cells are adaptive Treg cells generated in the periphery including IL-10 producing Tr1 cells (37). It will be important to characterize FCRL3 expression in this population. In preliminary experiments, we did not detect a change of FCRL3 expression in our attempts to expand IL-10 producing Treg cells by polyclonal activation of human CD4+ T cells in the presence of dexamethasone and vitamin D3 (49).
The FCRL3 gene is present in humans but not in mice (1, 2, 5). Genome sequence analysis suggests that the repertoire of FCRLs was shaped in evolution by an extensive recombination process including species-specific gain and loss of distinct exons or entire genes. This has resulted in remarkable diversity of surface FCRL proteins among species. There are only FCRL1, 5 and 6 in mice whereas humans possess FCRL1, 2, 3, 4, 5 and 6. Differences between human nTreg cells and mouse nTreg cells have been repeatedly reported, especially as to CD25 expression level and Foxp3 expression (20, 22, 23, 46). FCRL3 might be associated with human-specific immunoregulation of nTreg cells. The biological role of FCRL3 and its natural ligand are unknown but the gene locus 1q21–q23 where the FCRL3 locates has been implicated in susceptibility to autoimmune diseases (50). Recent studies suggest that the up-regulation of FCRL3 transcription induced by single nucleotide polymorphisms in the promoter of FCRL3 is associated with predisposition towards several autoimmune diseases (8, 50, 51). Although the link between FCRL3 and autoimmune diseases is not clear, it will be of importance to survey FCRL3 expression on nTreg cells from patient lymphocytes in the future.
We thank Drs. Ethan M. Shevach, Jay A. Berzofsky, Thomas A. Waldmann, Kenneth E. Santora and W. Keith Miskimins for their helpful comments on the manuscript. We thank Dr. Steven A. Rosenberg for tumor infiltrating lymphocytes. We thank Ms. Barbara J. Taylor for the critical assistance in flow cytometry and cell sorting, and Ms. Cathy Christopherson for editorial assistance.
The authors have no conflicting financial interest.
1This research was supported by the Intramural Research Program of the NIH, National Cancer Institute, Center for Cancer Research.
4Abbreviations used in this paper: CTLA-4, cytotoxic T-lymphocyte antigen 4; FcRH, Fc receptor homolog; FCRL, Fc receptor-like; Foxp3, forkhead box p3; GITR, glucocorticoid-induced TNF receptor family related protein; IFGP, Ig-Fc-gp42 related genes; IRTA, immune receptor translocation-associated; nTreg, naturally occurring CD4+ regulatory T; SPAP, SH2 domain-containing phosphatase anchor proteins; Teff, T effector; Treg, regulatory T
5The online version of this article contains supplemental material.