Latency-Associated Peptide Identifies a Novel CD4⁺CD25⁺Regulatory T Cell Subset with TGFβ-Mediated Function and Enhanced Suppression of Experimental Autoimmune Encephalomyelitis

Abstract

CD4⁺CD25⁺ regulatory T cells (Tregs) are essential for maintaining self-tolerance and immune homeostasis. Here we characterize a novel subset of CD4⁺CD25⁺ Tregs that express latency-associated peptide (LAP) on their cell surface (CD4⁺CD25⁺LAP⁺ cells). CD4⁺CD25⁺LAP⁺ cells express elevated levels of Foxp3 and Treg-associated molecules (CTLA4, glucocorticoid-induced TNFR-related gene), secrete TGFβ, and express both cell surface TGFβ and surface receptors for TGFβ. In vitro, the suppressive function of CD4⁺CD25⁺LAP⁺ cells is both cell contact and soluble factor dependent; this contrasts with CD4⁺CD25⁺LAP⁻ cells, which are mainly cell contact dependent. In a model of experimental autoimmune encephalomyelitis, CD4⁺CD25⁺LAP⁺ cells exhibit more potent suppressive activity than CD4⁺CD25⁺LAP⁻ cells, and the suppression is TGFβ dependent. We further show that CD4⁺CD25⁺LAP⁺ cells suppress myelin oligodendrocyte glycoprotein-specific immune responses by inducing Foxp3 and by inhibiting IL-17 production. Our findings demonstrate that CD4⁺CD25⁺ Tregs are a heterogeneous population and that the CD4⁺CD25⁺ subset that expresses LAP functions in a TGFβ-dependent manner and has greater in vivo suppressive properties. Our work helps elucidate the ambiguity concerning the role of TGFβ in CD4⁺CD25⁺ Treg-mediated suppression and indicates that LAP is an authentic marker able to identify a TGFβ-expressing CD4⁺CD25⁺ Treg subset.

How the immune system protects the host from pathogens while controlling unwanted or aberrant immune responses harmful to the host is a fundamental issue in immunology. Accumulating evidence suggests that a specialized subpopulation of T cells, the regulatory T cells (Tregs),³ actively suppress pathological and physiological immune responses and contribute to maintenance of immunological self-tolerance and immune homeostasis (1-3). Most Tregs constitutively express the IL-2R α-chain (CD25), and their development and function depend on the expression of transcription factor forkhead box P3 (Foxp3) (1, 4, 5).

A variety of mechanisms have been reported concerning how CD4⁺CD25⁺ Tregs suppress immune responses at the molecular level. However, no single mechanism appears to account for all features of suppression, and it is likely that CD4⁺CD25⁺ Tregs may regulate immunity by more than one mechanism. In rodents and humans, suppression mediated by CD4⁺CD25⁺ Tregs in vitro is dependent on a cell-to-cell contact, but cytokine-independent mechanism (6 - 8). However, several in vivo studies support the role of IL-10 in Treg suppression. It has been demonstrated that IL-10 is required for the homeostatic maintenance of the T cell number by Tregs (9) and is involved in Treg-mediated suppression in murine models of transplantation, graft-vs-host disease, chronic parasite infection, colitis, and a rat model of type 1 diabetes (10). The function of TGFβ in Treg suppression is controversial (3). Failure of anti-TGFβ to abrogate in vitro suppression as well as the fact that Tregs isolated from neonatal TGFβ knockout mice exhibit normal suppressive activity indicated that the in vitro suppressive capacity of CD4⁺CD25⁺ Tregs does not require TGFβ (1, 8). Nevertheless, several lines of evidence suggested that TGFβ may not necessarily act as a soluble factor but is expressed on the surface of activated Tregs and therefore functions in a membrane-proximal manner (11, 12). In vivo, anti-TGFβ treatment abrogated Treg suppression in a mouse model of colitis (13). Furthermore, expression of a T cell-specific dominant negative form of the TGFβ receptor II elicited inflammatory bowel disease in mice (14), and the suppression of CD8⁺ T cells that mediate autoimmunity (15) or tumor rejection (16) by CD4⁺CD25⁺ Tregs required an intact TGFβ receptor II on the CD8⁺ T cells. Nonetheless, others reported that CD4⁺CD25⁺ Tregs isolated from TGFβ⁻ mice could prevent inflammatory bowel disease in vivo (8). One possible explanation for the conflicting results regarding the contribution of TGFβ to CD4⁺CD25⁺ Treg-mediated suppression is that different subpopulations of CD4⁺CD25⁺ Tregs exist so that some can produce distinct immunosuppressive cytokines, and others can suppress solely by cell contact-dependent mechanisms.

Various T cell subpopulations, which are developmentally, phenotypically, or functionally different, have been purported to possess regulatory activity (17, 18). We have previously identified Th3 cells as TGFβ-secreting Tregs, which are preferentially generated via oral administration of autoantigens and are functionally mediated by TGFβ (19, 20). More recently, we and others have further identified TGFβ-dependent Tregs characterized by surface expression of latency-associated peptide (LAP; Refs. 21 and 22), which is the N-terminal propeptide of TGFβ precursor peptide. LAP remains noncovalently associated with TGFβ after cleavage from TGFβ precursor peptide by specific protease and forms the inactive latent TGFβ complex; it therefore contributes to the prevention of uncontrolled activation of the cognate TGFβ receptors (23, 24). CD4⁺LAP⁺ cells are active in the animal model of colitis (21, 22). Furthermore, oral administration of CD3-specific Ab suppresses experimental autoimmune encephalomyelitis (EAE, an animal model for multiple sclerosis; see Ref. 25) and acts by inducing TGFβ-dependent CD4⁺CD25⁻LAP⁺ T cells (26). In this study, we identified and characterized a subpopulation of CD4⁺CD25⁺ Tregs that express LAP on the surface and are potent regulators of EAE (CD4⁺CD25⁺LAP⁺ cells). LAP-expressing CD4⁺CD25⁺ Tregs are endowed with more potent suppressive activity and function in a TGFβ-dependent manner.

Materials and Methods

Mice

Female SJL and C57BL/6 (B6) mice at 6–8 wk of age were purchased from The Jackson Laboratory. Myelin oligodendrocyte glycoprotein (MOG) TCR-transgenic (Tg; 2D2) mice were kindly provided by Dr. V. K. Kuchroo (Center for Neurologic Diseases, Brigham and Women's Hospital, Boston, MA) and have been described (27). Mice were kept in a conventional, specific pathogen-free facility at the Harvard Institutes of Medicine according to the animal protocol guidelines of the Committee on Animals of Harvard Medical School (Boston, MA).

Induction and assessment of EAE

Female SJL mice were immunized s.c. in the flanks with 50 μg of proteolipid protein (PLP)_139–151 (HSLGKWLGHPDKF) in CFA containing 250 μg of Mycobacterium tuberculosis H37RA (Difco), followed by i.p. injection of 100 ng of pertussis toxin (List Biological Laboratories). Clinical assessment of EAE was performed according to the following criteria: 0, no signs of disease; 1, loss of tone in the tail; 2, hind limb weakness or partial paralysis; 3, complete hind limb paralysis; 4, front and hind limb paralysis; 5, moribund state. In some experiments, the indicated number and sorted populations of cells were adoptively transferred i.v. 2 days before immunization. All experiments were conducted in accordance with guidelines prescribed by the Institutional Animal Care and Use Committee at Harvard Medical School.

Abs and FACS analysis

CD16-CD32-specific Ab (FcBlock), fluorescein-conjugated mAbs to CD4 (GK1.5), PE-conjugated mAbs to CD4 (GK1.5), CD25 (PC61), IL-2 (JES6-5H4), IFN-γ (XMG.1.2), CTLA4 (UC10-4F10-11), CD103 (M290), CD45RB (16A), CD90.1 (OX-7), CD28 (37.51), rat IgG1 isotype control (R3-34), rat IgG2b isotype control (A95-1), Armenian hamster IgG isotype control and Syrian hamster IgG isotype control, streptavidin-PE (Sav-PE), PerCP-conjugated mAbs to CD4 (RM4-5), CD90.1 (OX-7), PerCP-Cy5.5-conjugated mAbs CD25 (PC61), allophycocyanin-conjugated mAbs to CD4 (RM4-5), CD25 (PC61), streptavidin-allophycocyanin (Sav-APC), and normal mouse IgG were purchased from BD Biosciences. PE-conjugated anti-mouse glucocorticoid-induced TNFR-related gene (GITR)/TNFRSF18, affinity-purified biotinylated goat anti-LAP polyclonal Ab, recombinant human LAP, anti-TGFβ1, −2, and −3 mAbs (1D11), biotinylated anti-TGFβRI, anti-TGFβRII, biotinylated chicken anti-TGFβ Ab (a polyclonal Ab recognizes epitopes that reside in active TGFβ1), mouse anti-LAP mAb 27232.11, biotinylated normal goat IgG, and normal goat IgG were purchased from R&D Systems. Biotinylated chicken IgY, peroxidase-conjugated streptavidin, and peroxidase-conjugated goat anti-mouse IgG (H + L) were purchased from Jackson ImmunoResearch Laboratories. PE-conjugated anti-mouse Foxp3 (clone FJK-16s), mAbs to ICOS (C398.4A), OX40 (OX-86), PD1 (J43), Tim3 (8B.2C12), and rat IgG2a isotype control were purchased from eBioscience. Anti-actin mAb (AC-40) was purchased from Sigma-Aldrich. Surface stainings were performed according to standard procedures at a density of 1–2 × 10⁶ cells per 50 μl, and volumes were scaled up accordingly. Flow cytometric analysis was performed on a FACSCalibur (BD Biosciences) with the use of FlowJo software (BD Biosciences). Foxp3 staining and analysis were performed by flow cytometry using the Foxp3 staining set (clone FJK-16s; eBioscience) according to the manufacturer's instructions.

Immunization

Mice were immunized s.c. with 50 μg of peptide MOG_35–55 (MEVGW YRSPFSRVVHLYRNGK) emulsified in CFA (Difco Bacto) containing 200 μgof M. tuberculosis H37RA (Difco). At the indicated time points, the animals were sacrificed, and draining lymph nodes were harvested for analysis.

Cytokine assay

All cytokines except TGFβ1 were measured by a multiplex Luminex assay (Upstate) according to the manufacturer's instructions. Briefly, cell culture supernatants were incubated with a suspension of analyte capture Ab-conjugated microspheres. After further incubation with biotinylated detection Abs and PE-conjugated streptavidin, fluorescent signal was read on a Luminex 100 system (Applied Cytometry Systems). TGFβ1 was measured by the Quantikine ELISA Kit (R&D Systems) with acidification according to the manufacturer's instruction.

Transwell assay

Transwell experiments were conducted to investigate the role of cell contact in the mechanism of suppression. Once sorted, indicated regulatory populations were cultured either in the lower chambers directly in contact with the responder CD4⁺CD25⁻LAP⁻cells or in the upper chambers separated from the responder cells by a 0.4-μm pore size membrane (Corning), which allows diffusion of small molecules, such as cytokines, but not of cells. Cells were stimulated with 1 μg/ml anti-CD3 in the presence of irradiated (3000 rad) syngeneic splenic APCs. The proliferation of cells from the lower chambers was measured by scintillation counting after pulsing with 1 μCi/well [³H]thymidine (1 Ci = 37 GBq) for the last 16 h of a 72-h incubation period.

CFSE labeling

For in vitro suppression assay, sorted CD4⁺CD25⁻LAP⁻ (10⁷ cells/ml) were incubated for 10 min at 37°C with 1 μM CFSE (Molecular Probes); for in vivo adoptive cotransfer experiments, CD25-depleted MOG TCR Tg T cells were labeled with CFSE (10 μM) at 37°C for 10 min.

Small interfering RNA (SiRNA) knockdown of TGFβ1

All SiRNA duplexes were obtained from Qiagen-Xeragon. SiRNA transfection was performed according to the manufacturer's instructions. RNA from transfected cells was used for real-time PCR to validate knockdown of the target genes.

Real-time PCR

The expression of TGFβ1 was determined using specific primers and probes (Applied Biosystems). TGFβ1 expression was normalized to the expression of the housekeeping gene β-actin.

Intracellular cytokine staining

Culture cells or cells from the draining lymph nodes of immunized animals were stimulated in culture medium containing PMA (20 ng/ml; Sigma-Aldrich), ionomycin (250 ng/ml; Sigma-Aldrich), and monensin (GolgiStop, 1 μl/ml; BD Biosciences), and cultures were incubated at 37°C in a humidified 5% CO₂ atmosphere for 4–6 h. Cells were harvested and incubated with 2.4G2 FcR-blocking Ab and then stained for surface markers. After staining of surface markers, cells were fixed and permeabilized using Cytofix/Cytoperm and Perm/Wash buffer (BD Biosciences) according to the manufacturer's instructions. Cells were then incubated at 4°C for 30 min with indicated Abs to cytokines and corresponding isotype controls and washed twice in Perm/Wash buffer before analysis.

Purification and adoptive transfer of cells

Pooled cells from spleens and peripheral lymph nodes (mesenteric, axillary, popliteal, inguinal, and cervical) of female SJL or B6 mice (6–10 wk) were subjected to erythrocyte lysis. After incubation with FcR-blocking Ab 2.4G2, cells were incubated with anti-CD4 microbeads (Miltenyi Biotec), and CD4⁺ cells were positively selected on LS MACS columns (Miltenyi Biotec), routinely achieving purities of <95%. Purified CD4⁺ cells then were stained with biotinylated LAP-specific Ab followed by streptavidin-PE, anti-CD25 allophycocyanin, and fluorescein-labeled anti-CD4. CD4⁺CD25⁺LAP⁺, CD4⁺CD25⁺LAP⁻, CD4⁺CD25⁻LAP⁺, and CD4⁺CD25⁻LAP⁻ cells were further sorted by using a FACSAria cell sorter (BD Biosciences). The purity of each population was <98% by FACS analysis. To purify CD25-depleted MOG TCR Tg T cells, pooled cells from spleens and lymph nodes of MOG TCR Tg mice were stained with biotinylated anti-CD25 followed by streptavidin microbeads; after depletion of CD25⁺ cells by LD MACS column (Miltenyi Biotec), CD4⁺ cells in the CD25⁻ flowthrough fraction were further purified by staining with CD4 microbeads and separated by MS or LS MACS columns (Miltenyi Biotec). Cells were injected into the lateral tail vein in a volume of 200 μl of PBS. Where indicated, cells were labeled with CFSE (Molecular Probes) by incubation for 10 min at 37°C in 10 μM CFSE in PBS, 0.1% BSA at a density of 1 × 10⁷ cells/ml.

Proliferation assays

For suppression assays, 1 × 10⁵ sorted CD4⁺CD25⁻LAP⁺, CD4⁺CD25⁺LAP⁺, CD4⁺CD25⁻LAP⁻, or CD4⁺CD25⁺LAP⁻ cells were cultured at a 1:1 ratio with syngeneic CD4⁺CD25⁻LAP⁻ cells. Cells were stimulated with anti-CD3 Ab (1 μg/ml) in the presence of irradiated (3000 rad) syngeneic splenic APCs in 200 μl of RPMI 1640 supplemented with 10% FCS in 96-well round-bottom plates. Proliferation was measured by scintillation counting after pulsing with 1 μCi per well [³H]thymidine for the last 16 h of a 72-h incubation period.

Similar assays were performed by using CFSE-labeled CD4⁺CD25⁻LAP⁻ responder cells, 2 × 10⁴ sorted CD4⁺CD25⁻LAP⁺, CD4⁺CD25⁺LAP⁺, CD4⁺CD25⁻LAP⁻, or CD4⁺CD25⁺LAP⁻ cells were cultured at a 1:1 ratio with CFSE-labeled CD4⁺CD25⁻LAP⁻ cells and were stimulated with anti-CD3 and APCs. After 3 days, the proliferation of responder cells was analyzed by FACS.

In vivo neutralization of TGFβ

For in vivo neutralization of TGFβ, SJL mice were immunized with 50 μg of PLP_139–151 2 days after adoptive transfer. Mice received five i.p. injections of anti-mouse TGFβ (clone 1D11; BioExpress) or isotype control on alternating days beginning 1 day post-adoptive transfer.

Immunofluorescence

Frozen sections (10 μm) were blocked for 1 h followed by incubation overnight with anti-CD4 Ab. Sections were then incubated with goat anti-rat Alexa Fluor 594 for 2 h at room temperature after being washed three times.

Immunoblot analysis

To obtain cell lysates, LAP⁺ and LAP⁻ T cells sorted from pooled spleens and lymph nodes of naive SJL mice were lysed in lysis buffer supplemented with protease inhibitor mixtures (Roche Molecular Biochemicals) for 10 min on ice and centrifuged at 14,000 × g for 15 min at 4°C; and supernatants were collected. The lysates were mixed with NuPAGE lithium dodecyl sulfate sample buffer (Invitrogen), incubated at 70°C for 10 min, run in 4–12% NuPAGE Bis-Tris gel (Invitrogen), and transferred to a polyvinylidene difluoride membrane (Pierce Biotechnology). The blotted membrane was blocked with 1% BSA, TBS, 0.05% Tween 20; washed; and incubated with 0.2 μg/ml biotin-conjugated anti-TGFβR Abs (R&D Systems) or purified anti-mouse actin Ab (Sigma-Aldrich; clone AC-40). After being washed, the membrane was incubated with HRP-conjugated secondary reagents. Then, the membrane was developed by SuperSignal West Dura Extended Substrate (Pierce Biotechnology) and exposed to an x-ray film after being washed.

Statistical analysis

Statistical significance was assessed by the two-tailed Student t test. For in vivo EAE experiments, differences in clinical scores were analyzed using one-way ANOVA, followed by Tukey multiple comparisons. p values of >0.05 were regarded as significant.

Results

Elevated expression of Treg-associated molecules of CD4⁺CD25⁺LAP⁺ cells

We and others have shown that CD4⁺LAP⁺ cells actively suppress autoimmune diseases in animal models of colitis and EAE and function in a TGFβ-dependent manner (21, 26, 28). Here we characterize the regulatory function of a small fraction of CD4⁺ cells that coexpress CD25 and LAP on the cell surface (CD4⁺CD25⁺LAP⁺ cells). Fig. 1A shows that CD4⁺LAP⁺ cells comprise two subsets, CD4⁺CD25⁻LAP⁺ and CD4⁺CD25⁺LAP⁺ cells. These two subsets consist of ~1.2% (CD25⁻LAP⁺) and >0.3% (CD25⁺LAP⁺) of CD4⁺ cells pooled from spleen and lymph nodes of naive SJL mice. In contrast, the fraction of CD4⁺CD25⁺LAP⁻ cells was substantially higher (5.5%) among CD4⁺ cells. Because CD4⁺CD25⁺LAP⁺ cells expressed CD25, we first determined whether they expressed signature Treg-associated molecules and exhibit the characteristic phenotype of Tregs.

FIGURE 1.

Phenotypic characterization of CD4⁺ populations sorted by CD25 and LAP expression. A, Frequencies of CD4⁺ populations analyzed by CD25 and LAP cell surface markers. CD4⁺ cells isolated from spleens and lymph nodes of naive SJL mice were stained with CD25- and LAP-specific Abs and analyzed by FACS. B, Expression of Treg-associated molecules. C, Expression of costimulatory molecules and Tim3. Flow cytometry of expression of Foxp3 (assessed by intracellular staining), GITR, intracellular CTLA4, CD103, and CD45RB (B), CD28, ICOS, OX40, PD1, and Tim3 (C), by CD4⁺CD25⁻LAP⁺, CD4⁺CD25⁺LAP⁺, CD4⁺CD25⁻LAP⁻, and CD4⁺CD25⁺LAP⁻ cells from pooled spleens and lymph nodes of naive SJL mice. Numbers next to outlined areas in B indicate percentage of cells positive for marker among each population. C, Red line, isotype-matched control IgG staining; blue line, anti-costimulatory molecules or anti-Tim3 staining. Data are representative of three independent experiments.

Foxp3 is a key molecular marker of CD4⁺CD25⁺ Tregs (4, 5). We found that Foxp3 expression (mean fluorescence intensity; MFI) of CD4⁺CD25⁺LAP⁺ cells was higher than the other three CD4⁺ subsets defined by CD25 and LAP expression including CD25⁺LAP⁻ cells (MFI 350 vs 247; Fig. 1B). CTLA4 has been shown to play an important role in the regulatory function of CD4⁺CD25⁺ cells both in vitro and in vivo (29). In addition, GITR (TNFRSF18), a receptor belonging to the TNFR superfamily (TNFRSF), is expressed constitutively at high levels and further up-regulated upon activation in CD4⁺CD25⁺ Tregs (30, 31). We thus examined the expression of CTLA4 and GITR of the four CD4⁺ subpopulations. As shown in Fig. 1B, CD4⁺CD25⁺LAP⁺ cells showed markedly higher expression of CTLA4 and GITR compared with CD25⁺LAP⁻ cells as well as the other two subsets. CD103 is a target of Foxp3 (5) and expressed on CD4⁺ Tregs (32, 33). Both the percentage of CD103-expressing CD25⁺LAP⁺ cells and the expression level (MFI) of CD103 on CD25⁺LAP⁺ cells were higher than CD25⁺LAP⁻ cells as well as the other two sub-populations (Fig. 1B). CD45RB^low is a phenotypic marker of regulatory T cells, and CD4⁺CD25⁻CD45RB^low cells suppress colitis in animal models (21, 34). The expression of CD45RB on CD4⁺CD25⁺LAP⁺ cells was decreased to a similar level as CD4⁺CD25⁺LAP⁻ cells. Finally, CD4⁺CD25⁺LAP⁺ cells were present in all lymphoid compartments examined including the thymus, and there appeared to be less CD4⁺CD25⁺LAP⁺ cells in mesenteric lymph node (not shown).

Elevated expression of ICOS, OX40, PD1, and Tim3 on CD4⁺CD25⁺LAP⁺ cells

Costimulatory molecules expressed by Tregs are involved in their function (35) and the Tim3 pathway plays an important role in down-regulating Th1 responses (36) and facilitating the development of immunological tolerance. As shown in Fig. 1C, among the four populations examined, CD4⁺CD25⁺LAP⁺ cells expressed the highest levels of costimulatory molecules and Tim3. Furthermore, expression of LAP was associated with up-regulation of ICOS, PD1, and Tim3, given that both CD4⁺ CD25⁻LAP⁺ and CD4⁺CD25⁺LAP⁺ cells expressed higher levels of the aforementioned cell surface molecules than LAP⁻ populations. The expression level of CD28 was similar between four populations.

CD4⁺CD25⁺LAP⁺ cells secrete suppressive cytokine TGFβ

To further address the function of CD4⁺CD25⁺LAP⁺cells, cytokine production of the four CD4⁺ subpopulations aforementioned was compared (Table I). The cytokine profiles for GM-CSF, IFN-γ, IL-12, IL-6, and TNF-α were not significantly different between both CD4⁺CD25⁺ subpopulations (CD4⁺CD25⁺LAP⁺ and CD4⁺CD25⁺LAP⁻); however, CD4⁺CD25⁺LAP⁺ produced more IL-2, IL-10, and IL-4 than CD4⁺CD25⁺LAP⁻ cells, and distinct from CD4⁺CD25⁺LAP⁻ cells, CD4⁺CD25⁺LAP⁺ cells produced TGFβ (p = 0.0017). Although the cytokine profiles for GM-CSF, IFN-γ, IL-12, IL-6, and TNF-α were not significantly different between the two CD4⁺CD25⁻ subsets (CD4⁺CD25⁻LAP⁺ and CD4⁺CD25⁻LAP⁻), CD4⁺CD25⁻LAP⁺ cells secreted significantly higher amounts of IL-10 (p = 0.0314), IL-4 (p = 0.0008), IL-5 (p = 0.0465) than CD4⁺CD25⁻LAP⁻ cells, and unlike CD4⁺CD25⁻LAP⁻ cells, CD4⁺CD25⁻LAP⁺ cells secreted TGFβ (p = 0.0054). Only the subpopulations that express LAP secreted TGFβ.

Table I.

Cytokine production of CD4⁺ subpopulations sorted by CD25 and LAP expression^a

Cytokine	GM-CSF (pg/ml)	IFN-γ (pg/ml)	IL-12 (pg/ml)	IL-6 (pg/ml)	TNF-α (pg/ml)	IL-2 (pg/ml)	IL-10 (pg/ml)	IL-4 (pg/ml)	IL-5 (pg/ml)	TGFβ (pg/ml)
CD4+CD25−LAP−	6950.0 ± 403.1	16790.0 ± 1732.4	33.8 ± 14.3	29.4 ± 0.4	9860.0 ± 3648.7	17.0 ± 1.7	79.2 ± 6.4	144.3 ± 2.8	82.0 ± 0.4	0.0 ± 0.0
CD4+CD25−LAP+	3180.0 ± 226.3	9160.0 ± 792.0	145.7 ± 26.0	124.3 ± 8.3	1428.0 ± 316.8	7.5 ± 3.5	16470.0 ± 572.8	3160.0 ± 0.0	546.0 ± 24.0	1099.3 ± 13.2
CD4+CD25+LAP−	172.1 ± 8.1	732.0 ± 22.6	58.0 ± 1.0	1.9 ± 0.2	26.9 ± 1.2	0.0 ± 0.0	12.7 ± 2.7	2.3 ± 0.6	12.7 ± 2.3	0.0 ± 0.0
CD4+CD25+LAP+	109.0 ± 11.2	429.0 ± 57.3	56.0 ± 1.0	1.7 ± 0.3	50.4 ±11.0	129.1 ± 29.8	91.3 ± 19.6	14.0 ± 3.3	6.5 ± 4.2	325.1 ± 1.3

^aCD4⁺CD25⁻LAP⁻, CD4⁺CD25⁻LAP⁺, CD4⁺CD25⁺LAP⁻, and CD4⁺CD25⁺LAP⁺ cells were sorted from pooled spleens and lymph nodes of naive SJL mice as described in Materials and Methods. After sorting, 1 × 10⁵ cells from each sorted population were stimulated with plate-bound CD3-specific Ab (10 μ/ml). Cytokines were measured by ELISA or Luminex. Results are mean ± SD of triplicate wells and are representative of three independent experiments.

Regulatory function of CD4⁺CD25⁺LAP⁺ cells in vitro

Because CD4⁺CD25⁺LAP⁺ cells had phenotypic hallmarks of Tregs, we next investigated whether they exhibited functional characteristics of Tregs in vitro. CD4⁺CD25⁺LAP⁺ cells were markedly anergic relative to the other three populations (Fig. 2A). Given that the four CD4⁺ subpopulations had different anergic properties (Fig. 2A), we compared the suppressive capacity of these cell populations in coculture assays using CFSE-labeled responder cells (CD4⁺CD25⁻LAP⁻). As shown in Fig. 2B, CD4⁺CD25⁺LAP⁺ cells had the most potent in vitro suppressive activity, and the difference in suppression mediated by CD4⁺CD25⁺LAP⁻ cells and CD4⁺CD25⁺LAP⁺ subsets was observed in three replicate experiments (CD4⁺CD25⁺LAP⁻ vs CD4⁺CD25⁺LAP⁺: 69.3% vs 85.7%, 69.6% vs 84.7% and 70.3% vs 85.6%, p > 0.000004).

FIGURE 2.

Regulatory function of CD4⁺ subsets sorted by CD25 and LAP cell surface markers in vitro. A, Proliferation of CD4⁺ populations sorted by CD25 and LAP cell surface markers. Sorted CD4⁺CD25⁻LAP⁺, CD4⁺CD25⁺LAP⁺, CD4⁺CD25⁻LAP⁻, or CD4⁺CD25⁺LAP⁻ cells (1 × 10⁵) were cultured together with irradiated (3000 rads) syngeneic splenocytes in the presence of 1 μg/ml anti-CD3, and proliferation was assessed by scintillation counting after a pulse with [³H]thymidine for the last 16 h of a 72-h incubation period. Data are presented as means ± 95% confidence intervals for the mean. B, Suppressive function of CD4⁺ populations sorted by CD25 and LAP cell surface markers. The proliferation of CFSE-labeled responder CD4⁺ CD25⁻LAP⁻ cells was analyzed by FACS after being cultured with sorted CD4⁺CD25⁻LAP⁺, CD4⁺CD25⁺LAP⁺, CD4⁺CD25⁻LAP⁻, or CD4⁺CD25⁺LAP⁻ cells for 72 h. C, Transwell assay. Responder CD4⁺CD25⁻LAP⁻ cells (1 × 10⁵) were stimulated with anti-CD3 alone or in the presence of an equal number of sorted CD4⁺CD25⁻LAP⁺, CD4⁺CD25⁺LAP⁺, CD4⁺CD25⁻LAP⁻, or CD4⁺CD25⁺LAP⁻ cells, either in direct contact or separated by a transwell inset (Transwell) as described in Materials and Methods. Number next to each bar, percentage of suppression. Data are representative of at least two independent experiments and are presented as means ± 95% confidence intervals for the mean. D, IFN-γ production of responder CD4⁺CD25⁻LAP⁻ cells in vitro. CD4⁺CD25⁻LAP⁻ responder cells were cultured with CD4⁺CD25⁺LAP⁺ or CD4⁺CD25⁺LAP⁻ populations in the presence of anti-CD3 Ab (1 μg/ml) and irradiated (3000 rad) syngeneic splenic APCs for 60 h, IFN-γ production by responder cells were then determined by intracellular cytokine staining. Data were expressed as the percent suppression of IFN-γ production.

To assess whether CD4⁺CD25⁺LAP⁺ cells exerted their regulatory function through direct cell contact or release of soluble factors, we conducted transwell experiments. Both CD4⁺CD25⁺LAP⁺ and CD4⁺CD25⁺LAP⁻ cells markedly inhibited the proliferation of responder cells. CD4⁺CD25⁺LAP⁻ cells exerted their suppression mainly through cell contact; whereas the suppressive function of CD4⁺CD25⁺LAP⁺ cells was both cell contact and soluble factor dependent given that the suppression could not be completely reversed in the transwell assay (Fig. 2C). To further investigate the regulatory function of CD4⁺CD25⁺LAP⁺ cells in vitro, we examined whether CD4⁺CD25⁺LAP⁺ cells affected the production of IFN-γ by responder cells. As shown in Fig. 2D, CD4⁺CD25⁺LAP⁺ cells suppressed IFN-γ production by responder cells to a greater extent than did CD4⁺CD25⁺LAP⁻ cells (p > 0.0004). The production of IL-17 was undetectable in the in vitro assay (not shown). These results demonstrate that CD4⁺CD25⁺LAP⁺ cells have different in vitro suppressive properties from CD4⁺CD25⁺LAP⁻ cells.

CD4⁺CD25⁺LAP⁺ cells express TGFβ1 and TGFβR on the cell surface

We next determined whether CD4⁺CD25⁺LAP⁺ cells expressed active TGFβ1 on the cell surface. As shown in Fig. 3A, virtually no CD4⁺CD25⁺LAP⁻ cells were cell surface TGFβ1 (mTGFβ) positive; however, ~10% of freshly isolated CD4⁺CD25⁺LAP⁺ cells expressed active TGFβ1 on the cell surface. Thus, CD4⁺CD25⁺LAP⁺ cells expressed not only latent TGFβ1 but also active TGFβ1 on the cell surface. This observation is consistent with our finding that CD4⁺CD25⁺LAP⁺ cells may mediate suppression via a cell contact-dependent mechanism (Fig. 2C). When we examined whether CD4⁺CD25⁺LAP⁺ cells expressed receptors for TGFβ (TGFβR) by flow cytometry analysis, we found that virtually all CD4⁺CD25⁺LAP⁻ cells were negative for both TGFβRI and TGFβRII, whereas a large fraction of CD4⁺CD25⁺LAP⁺ cells expressed TGFβRI and TGFβRII (Fig. 3, B and C). We then performed immunoblot analysis to determine whether LAP⁻ cells truly do not express receptors for TGFβ. As shown in Fig. 3D, both LAP⁺ and LAP⁻ populations expressed TGFβRs although LAP⁺ cells expressed slightly higher amounts (36%) of TGFβRII than LAP⁻ cells in the immunoblot analysis. These results suggest that there are TGFβ receptors on the surface of LAP⁻ cells, but they are masked and not accessible to measurement by Abs in flow cytometry analysis.

FIGURE 3.

Expression patterns of cell surface-bound TGFβ1 and TGFβRs on four CD4⁺ cell populations. CD4⁺ T cells from pooled spleens and lymph nodes of naive SJL mice were stained with CD25, LAP, and TGFβ1 (R&D Systems; chicken IgY) or TGFβR-specific Abs, the expression of cell surface-bound TGFβ1 (mTGFβ; A), TGFβRI (B), and TGFβRII (C) on each population was analyzed by FACS. Red line, iso-type-matched control IgG staining; blue line, anti-mTGFβ or TGFβR staining. Data are representative of at least two independent experiments. D, Immunoblot analysis for the expression of TGFβRs and actin of LAP⁺ and LAP⁻ cells. Cell lysates purified from sorted LAP⁺ and LAP⁻ T cells were subjected to SDS-PAGE and immunoblot analysis.

Adoptive transfer of CD4⁺CD25⁺LAP⁺ but not CD4⁺CD25⁺LAP⁻ cells efficiently ameliorate EAE

To investigate the function of CD4⁺CD25⁺LAP⁺ cells in vivo, we assessed the capacity of CD4⁺CD25⁺LAP⁺ cells to suppress EAE. CD4⁺CD25⁺LAP⁺ and CD4⁺CD25⁺LAP⁻ cells were adoptively transferred to SJL mice that were then immunized with PLP_139–151 to induce EAE 2 days after adoptive transfer (Fig. 4A). As shown in Fig. 4B and Table II, adoptive transfer of CD4⁺CD25⁺LAP⁺ cells suppressed EAE as measured by mean day of onset and mean maximal disease score. There was only a mild suppressive effect on the clinical course when CD4⁺CD25⁺LAP⁻ cells were transferred, and the effect was not maintained. These results demonstrate that CD4⁺CD25⁺LAP⁺ cells potently suppress EAE and do so more efficiently than CD4⁺CD25⁺LAP⁻ cells (p < 0.001, mean maximal scores 1.2 vs 2.5). In addition, immunofluorescence staining of spinal cords demonstrated that there was significantly reduced infiltration of CD4⁺ cells in the mouse group receiving CD4⁺CD25⁺LAP⁺ cells compared with control mice (150 vs 937 cells/mm²; p = 0.008, Fig. 4C).

FIGURE 4.

Effect of adoptive transfer of CD4⁺CD25⁺LAP⁺ and CD4⁺CD25⁺LAP⁻ cells on EAE. A, Schematic representation of experimental design. PT, Pertussis toxin. B, 1 × 10⁵ sorted CD4⁺CD25⁺LAP⁺ or CD4⁺CD25⁺LAP⁻ cells or PBS were injected i.v. into naive SJL mice. Mice were then immunized with PLP_139–151 in CFA to induce EAE 2 days after adoptive transfer. The mean daily score for each group (five mice per group) is shown. Data are representative of three experiments. C, Immunofluorescence microscopy of CD4⁺ cells in spinal cords. Cryosections of spinal cords from mice of control group (a) or mice receiving CD4⁺CD25⁺LAP⁺ cells (b) at the end of experiment were stained with anti-CD4. Results represent blinded analysis of three samples per group. ×400. Bar, 50 μm.

Table II.

EAE induced in SJL recipients of CD4⁺CD25⁺LAP⁺ and CD4⁺CD25⁺LAP⁻ cells

Treatmenta	No. of Mice/Group	Incidence	Mean Day of Onset	Mean Clinical Score on Day 10b	Mean Maximal Scorec
Control	5	5/5	8.5	2.9 ± 0.1d,e	2.9 ± 0.1f,g
CD4+CD25+LAP+	5	5/5	19.0	0.8 ± 0.6d	1.2 ± 0.7f
CD4+CD25+LAP−	5	5/5	11.0	1.9 ± 0.6e	2.5 ± 0.8g

^aSorted CD4⁺CD25⁺LAP⁺ or CD4⁺CD25⁺LAP⁻ cells (1 × 10⁵) were injected i.v. into naive SJL mice. Mice were then immunized with PLP_139–151 in CFA to induce EAE 2 days after adoptive transfer, and signs of EAE were monitored.

^bThe day when EAE of the control group reached peak.

^cData are the mean maximal clinical scores ± SEM.

^dp = 0.03.

^ep = 0.17.

^fp = 0.04.

^gp = 0.67.

CD4⁺CD25⁺LAP⁺ cells suppress IFN-γ production and induce Foxp3 expression of MOG TCR Tg T cells in vivo

To further elucidate the mechanism of suppression mediated by CD4⁺CD25⁺LAP⁺ cells in vivo, CD25-depleted, CFSE-labeled MOG TCR Tg Thy1.1⁺ T cells were adoptively cotransferred with CD4⁺CD25⁺LAP⁺ cells sorted from naive B6 mice or transferred alone into wild-type B6 (Thy1.2⁺) mice, and the effect of CD4⁺CD25⁺LAP⁺ cells on the response of MOG TCR Tg T cells (CD4⁺Thy1.1⁺) after immunization with MOG_35~55 in CFA was monitored (Fig. 5A). Proliferation of MOG TCR Tg T cells in draining lymph nodes was not significantly affected when mice were cotransferred with CD4⁺CD25⁺LAP⁺ cells (Fig. 5B). In addition, no differences in the frequency and activation (CD25 up-regulation) of MOG TCR Tg T cells in draining lymph nodes were observed between the mice receiving MOG TCR Tg T cells together with CD4⁺CD25⁺LAP⁺ cells and the mice receiving MOG TCR Tg T cells alone (not shown). Thus, the homing/expansion of MOG TCR Tg T cells was barely affected by CD4⁺CD25⁺LAP⁺ cells.

FIGURE 5.

Effect of adoptive cotransfer of CD4⁺CD25⁺LAP⁺ cells on the effector function of MOG TCR Tg T cells in vivo. A, Schematic representation of experimental design. CD25⁻, CFSE-labeled MOG TCR Tg Thy1.1⁺ T cells (3 × 10⁵) were transferred alone or together with CD4⁺CD25⁺LAP⁺ cells into B6 (Thy1.2⁺) mice. Two days later, recipients were immunized with 50 μg of MOG_35–55 peptide in CFA. Mice were killed 5 days after immunization, and cells from draining lymph nodes were harvested, stained, and analyzed by flow cytometry. B, Proliferation of adoptively transferred MOG TCR Tg T cells in the draining lymph nodes after immunization. The plots show the expression of CD4 vs CFSE fluorescence intensity on gated donor-derived cells (CD4⁺Thy1.1⁺). Numbers above boxed areas indicate the frequency of CFSE⁺ cells among CD4⁺Thy1.1⁺ cells. Frequencies are representative of three mice per group. C, IFN-γ production of transferred MOG TCR Tg T cells. Draining lymph node cells from mice adoptively transferred and immunized as described in A, were restimulated ex vivo with PMA-ionomycin and stained for intracellular IFN-γ. Numbers next to boxed areas indicate the frequency of IFN-γ⁺ cells among CD4⁺Thy1.1⁺ cells. D, Expression of Foxp3 of transferred MOG TCR Tg T cells. Draining lymph node cells from mice adoptively transferred and immunized as described above were stained for intracellular Foxp3. Numbers next to boxed areas indicate the frequency of Foxp3⁺ cells among CD4⁺Thy1.1⁺ cells. Frequencies are representative of three mice per group. Data are representative of at least two independent experiments.

We then measured cytokine production of MOG TCR Tg T cells in draining lymph nodes after restimulation in vitro five days after immunization. In recipients of MOG TCR Tg T cells alone, 15.3% of gated CD4⁺Thy1.1⁺ T cells produced IFN-γ. In contrast, the fraction of IFN-γ-producing MOG TCR Tg T cells was reduced (6.2%) in mice that had received CD4⁺CD25⁺LAP⁺ cells (Fig. 5C). IL-2 production of MOG TCR Tg T cells was similar between two groups of animals (not shown).

It has been shown that TGFβ can convert Foxp3⁻ T cells into Foxp3⁺ Tregs (37-39). Because the LAP molecule is closely associated with TGFβ, and CD4⁺CD25⁺LAP⁺ cells produced TGFβ and expressed mTGFβ (Table I and Fig. 3), we asked whether cotransfer of CD4⁺CD25⁺LAP⁺ cells could convert MOG TCR Tg T cells to Foxp3⁺ cells. Although CD4⁺CD25⁻ cells contain some Foxp3⁺ cells, virtually no Foxp3⁺ cells were detected among CD4⁺CD25⁻ fraction of MOG TCR Tg T cells from naive MOG TCR Tg mice (not shown). In addition, we depleted CD25⁺ cells from purified MOG TCR Tg T cells before adoptive cotransfer to insure that no Foxp3⁺ cells were being transferred. As shown in Fig. 5D, there was an increase in percentage of cells expressing Foxp3 among MOG TCR Tg T cells in draining lymph nodes of animals that had received both MOG TCR Tg T cells and CD4⁺CD25⁺LAP⁺ cells than in mice receiving MOG TCR Tg T cells alone (4.8% vs 1.6%; p = 0.039). Thus, CD4⁺CD25⁺LAP⁺ cells induce/expand Foxp3⁺ cells in vivo.

CD4⁺CD25⁺LAP⁺ cells suppress IL-17 production of MOG TCR Tg T cells in vivo

T cells producing IL-17 (Th17 cells) have been reported to be involved in autoimmune diseases such as EAE (40, 41). We thus investigated the effect of CD4⁺CD25⁺LAP⁺ cells on IL-17 production of MOG TCR Tg T cells according to the experimental design described above for Fig. 5. As shown in Fig. 6, a large fraction (62%) of MOG TCR Tg T cells in draining lymph nodes of mice receiving MOG TCR Tg T cells alone produced IL-17. Although the fraction of IL-17 producing MOG TCR Tg T cells was decreased by cotransfer of CD4⁺CD25⁺LAP⁻ cells (48% vs 62%, p = 0.16), CD4⁺CD25⁺LAP⁺ cells more efficiently suppressed IL-17 production of MOG TCR Tg T cells (30% vs 62%, p = 0.015).

FIGURE 6.

Regulation of ex vivo IL-17 production of MOG TCR Tg T cells by CD4⁺CD25⁺LAP⁺ cells and CD4⁺CD25⁺LAP⁻ cells. CD25⁻ and CFSE-labeled MOG TCR Tg Thy1.1⁺ T cells (3 × 10⁵) were transferred alone or together with CD4⁺CD25⁺LAP⁺ or CD4⁺CD25⁺LAP⁻ cells into B6 (Thy1.2⁺) mice. Two days later, recipients were immunized with 50 μg of MOG_35–55 peptide in CFA. Mice were killed 5 days after immunization, and cells from draining lymph nodes were harvested and stained for intracellular IL-17. Numbers next to boxed areas indicate the frequency of IL-17⁺ cells among CD4⁺Thy1.1⁺ cells. Frequencies are representative of three mice per group. Data are representative of at least two independent experiments.

Regulatory function of CD4⁺CD25⁺LAP⁺ cells is TGFβ dependent

To investigate the role of TGFβ in the suppressive function of CD4⁺CD25⁺LAP⁺ cells, we conducted RNA interference experiments to silence the expression of both secreted form and surface-bound TGFβ1 and studied the effect of TGFβ1 knockdown on the suppressive function of CD4⁺CD25⁺LAP⁺ cells in vitro. CD4⁺CD25⁺LAP⁺ cells that were mock-transfected or transfected with control SiRNA efficiently suppressed the proliferation of responder cells (Fig. 7A; p = 0.001), whereas TGFβ1 knockdown significantly reversed the suppressive function of CD4⁺CD25⁺LAP⁺ cells (Fig. 7A; p = 0.04). We then tested CD4⁺CD25⁺LAP⁻ cells. Although CD4⁺CD25⁺LAP⁻ cells were as suppressive as CD4⁺CD25⁺LAP⁺ cells in mock-transfected or control SiRNA transfected conditions, silencing TGFβ1 expression had no effect on the suppressive function of CD4⁺CD25⁺LAP⁻ cells (Fig. 7A).

FIGURE 7.

TGFβ-mediated suppressive function of CD4⁺CD25⁺LAP⁺ cells in vitro and in vivo. A, Effect of SiRNA knockdown of TGFβ1 on the in vitro suppressive function of CD4⁺CD25⁺LAP⁺ and CD4⁺CD25⁺LAP⁻ cells. Sorted CD4⁺CD25⁺LAP⁺ and CD4⁺CD25⁺LAP⁻ cells were transfected with 250 ng of SiRNA specific for TGFβ1. Negative controls consist of mock-transfected cells and cells transfected with control SiRNA. The knockdown extent of TGFβ1 siRNA was determined by real-time PCR and was 48 ± 15% and 34 ± 2% with reference to control siRNA transfected cells for CD4⁺CD25⁺LAP⁻ and CD4⁺CD25⁺LAP⁺ subsets, respectively. TGFβ1 expression was normalized to the expression of the housekeeping gene β-actin. Twenty-four hours after transfection, transfected cells were cultured at 1:1 ratio with responder CD4⁺CD25⁻LAP⁻ cells, and in vitro proliferation assays were performed as described in Materials and Methods. The results shown represent the mean ± SD of triplicate wells and are representative of two independent experiments. B, Effect of neutralization of TGFβ on the regulatory function of CD4⁺CD25⁺LAP⁺ cells in vivo. SJL mice were adoptively transferred with 1 × 10⁵ sorted CD4⁺CD25⁺LAP⁺ cells 2 days before EAE induction. Mice then received five i.p. injections of 50 μg of neutralizing TGFβ-specific or control Abs every other day starting 1 day before EAE induction. Values are the mean daily score for each group (five mice per group). Data are representative of at least two independent experiments.

To address the role of TGFβ in regulatory function of CD4⁺CD25⁺LAP⁺ cells in vivo in the EAE model, SJL mice receiving PBS or CD4⁺CD25⁺LAP⁺ cells were injected i.p. with five doses of neutralizing anti-TGFβ mAb (1D11) or control Ab starting one day before EAE induction (Fig. 7B). As shown in Fig. 7B, in vivo neutralization of TGFβ significantly reversed the suppression of EAE-mediated by CD4⁺CD25⁺LAP⁺ cells (Fig. 7B; p < 0.001). Adoptive transfer of CD4⁺CD25⁺LAP⁻ cells had only a minimal effect on EAE that was not influenced by anti-TGFβ (not shown). These results demonstrate that the function of CD4⁺CD25⁺LAP⁺ cells in vivo is TGFβ dependent.

Discussion

Here we have identified and characterized a unique subpopulation of CD4⁺CD25⁺ Tregs, which are CD4⁺CD25⁺LAP⁺ cells that coexpress CD25 and LAP on the cell surface and have more potent regulatory function in vivo than CD25⁺LAP⁻ cells. CD4⁺CD25⁺LAP⁺ cells express higher levels of the Treg-associated molecules Foxp3, GITR and CTLA4 compared with CD25⁺LAP⁻ cells as well as other CD4⁺ cell subpopulations sorted by CD25 and LAP expression. Like classical CD4⁺CD25⁺ Tregs, in vitro, CD4⁺CD25⁺LAP⁺ cells are anergic and suppressive; however, the in vitro suppressive function of CD4⁺CD25⁺LAP⁺ cells is only partially contact dependent, this contrasts with CD4⁺CD25⁺LAP⁻ cells, which are mainly cell contact dependent. In vivo, CD4⁺CD25⁺LAP⁺ cells suppressed EAE more efficiently than CD4⁺CD25⁺LAP⁻ cells. Furthermore, when CD4⁺CD25⁺LAP⁺ cells were cotransferred with MOG-specific TCR Tg T cells into mice, the CD4⁺CD25⁺LAP⁺ cells induced Foxp3 expression and suppressed IL-17 production by the MOG-specific TCR Tg T cells. Suppression of IL-17 appears to be more relevant to EAE pathogenesis than cell proliferation. Finally, CD4⁺CD25⁺LAP⁺ cells secrete TGFβ and express cell surface-bound TGFβ; furthermore, the regulatory function of CD4⁺CD25⁺LAP⁺ cells is TGFβ dependent in vitro and in vivo. Thus, LAP identifies a unique CD4⁺CD25⁺ Treg subpopulation that is linked to TGFβ and has greater suppressive properties than classical CD4⁺CD25⁺ Tregs.

The function of immunosuppressive cytokine TGFβ in Treg-mediated suppression remains controversial (3). Among the possible explanations for the discrepancies are that various subsets of Tregs exist and the function of each population is distinct in terms of TGFβ dependency (17, 42). In support of this possibility, we found that three of the four CD4⁺ subpopulations sorted by CD25 and LAP expression exhibited regulatory function and only the LAP⁺ subpopulations (CD4⁺CD25⁻LAP⁺ and CD4⁺CD25⁺LAP⁺) secreted TGFβ. Furthermore, both LAP⁺ subpopulations function in a TGFβ-dependent manner in an animal model of autoimmune encephalomyelitis as reported previously for CD4⁺CD25⁻LAP⁺ cells (26) and shown in this report for the CD4⁺CD25⁺LAP⁺ cells. Thus, our data established that the expression of LAP is associated with TGFβ dependency of Treg function. We have further demonstrated that CD4⁺CD25⁺LAP⁺ but not CD4⁺CD25⁺LAP⁻ cells express TGFβ in both cell surface-bound and soluble forms.

Although both CD4⁺CD25⁺LAP⁺ and CD4⁺CD25⁺LAP⁻ cell subsets showed regulatory function in vitro, and the latter makes up 95% of the Treg population, we demonstrated that CD4⁺CD25⁺LAP⁺ cells exhibited far greater suppressive activity in the EAE model compared with CD4⁺CD25⁺LAP⁻ cells, which only minimally suppressed EAE (Fig. 4). The fact that the suppression mediated by CD4⁺CD25⁺LAP⁻ cells was primarily cell contact dependent whereas the regulatory function of CD4⁺CD25⁺LAP⁺ cells was only partially dependent on cell contact is one possible explanation for the difference in in vivo suppressive properties between the two cell subsets. Not only can CD4⁺CD25⁺LAP⁺ cells suppress effector T cells in their close vicinity via cell contact, but also they possess soluble factor-dependent suppressive ability that renders this cell subset able to suppress effector T cells across a broader area and thus they are able to regulate autoimmunity more efficiently. Compared with CD4⁺CD25⁺LAP⁻ cells, CD4⁺CD25⁺LAP⁺ cells secrete immunosuppressive cytokine TGFβ, and this might represent an important mechanism by which CD4⁺CD25⁺LAP⁺ cells regulate effector T cells without cell-to-cell contact. Our finding that the function of CD4⁺CD25⁺LAP⁺ cells in vivo is TGFβ dependent provides support for this possibility (Fig. 7).

TGFβ plays an important role in the development, maintenance, induction, and expansion of Tregs and has been shown to convert naive T cells to CD4⁺CD25⁺ Tregs via the induction of Foxp3 expression (43, 44). The induction of regulatory activities by TGFβ correlates with the increased expression of Foxp3. In addition to secreting soluble TGFβ, CD4⁺CD25⁺LAP⁺ cells express TGFβ and TGFβRs on the cell surface. It is likely that the TGFβ produced/expressed by CD4⁺CD25⁺LAP⁺ cells is able to function in both autocrine and paracrine fashions for these cells. TGFβ secreted/expressed by a CD4⁺CD25⁺LAP⁺ cell can bind to TGFβR on the cell surface and maintain the peripheral homeostasis of the CD4⁺CD25⁺LAP⁺ cells or expand their pool size. In addition, the TGFβ produced or expressed by CD4⁺CD25⁺LAP⁺ cells may enhance their own Foxp3 expression and further promote their regulatory function. It has been shown that decreased Foxp3 expression in Tregs is associated with various immune diseases (45-47). More recently, Wan and Flavell (48) reported that decreased Foxp3 expression causes immune disorders by subverting the suppressive function of Tregs and converting Tregs into effector cells. Our results, in conjunction with aforementioned findings, suggest that TGFβ expressed by CD4⁺CD25⁺LAP⁺ cells might function via a self-amplifying positive autoregulatory loop in which TGFβ is responsible for the elevated Foxp3 expression observed for CD4⁺CD25⁺LAP⁺ cells. Foxp3 then further up-regulates the expression of signature genes for Tregs and down-regulates the expression of Smad7 that is induced by TGFβ and limits the TGFβ signaling (38), thus contributing to the enhanced regulatory function of CD4⁺CD25⁺LAP⁺ cells. Because cell surface-bound TGFβ and TGFβRs are expressed on freshly isolated CD4⁺CD25⁺LAP⁺ cells from naive mice, the TGFβ expressed on CD4⁺CD25⁺LAP⁺ cells is able to maintain or even enhance their regulatory function.

Although TGFβ plays an important role in CD4⁺CD25⁺LAP⁺-mediated suppression, our data cannot exclude the possibility that other molecules are also involved in the regulatory function of CD4⁺CD25⁺LAP⁺ cells. The observation that CD4⁺CD25⁺LAP⁺ cells express substantially higher levels of CTLA4 than CD4⁺CD25⁺ cells that are negative for LAP suggests that CTLA4 might be important for the function of CD4⁺CD25⁺LAP⁺ cells. It has been shown that CTLA4 expressed by natural Tregs has a key role in Treg-mediated suppression in vivo and in vitro (1). Moreover, CTLA4 signaling can facilitate the TGFβ-mediated suppression of CD4⁺CD25⁺ T cells and is required for TGFβ to induce Foxp3 and generate Tregs (49, 50). Whether CTLA4 is required for the function of CD4⁺CD25⁺LAP⁺ cells and, if so, whether it acts in synergy with TGFβ or has a distinct function remains to be determined.

In contrast with classical CD4⁺CD25⁺ natural Tregs the function of which relies on cell-to-cell contact, the suppressive function of CD4⁺CD25⁺LAP⁺ cells is dependent on both cell-to-cell contact and soluble factors as shown in a transwell assay. In addition, the regulatory function of CD4⁺CD25⁺LAP⁺ cells is TGFβ dependent; this contrasts with a number of studies suggesting that CD4⁺CD25⁺ natural Tregs suppress in a TGFβ-independent manner (2, 8). In our previous work, we have characterized CD4⁺CD25⁻LAP⁺ cells that are induced or expanded after oral administration of CD3-specific Ab and function in a TGFβ-dependent manner (26). Although CD4⁺CD25⁻LAP⁺ T cells produced higher levels of TGFβ than CD4⁺CD25⁺LAP⁺ T cells, the latter exhibited a stronger suppressive activity. We believe that this difference is related to the increased percentage of Foxp3⁺ cells and enhanced expression of Foxp3 in CD4⁺CD25⁺LAP⁺ cells and because CD4⁺CD25⁻LAP⁺ cells secreted Th1 and Th2 cytokines plus the proinflammatory cytokines IL-6 and TNF-α which were not secreted by CD4⁺CD25⁺LAP⁺ cells. It has been shown that TNF-α and IL-6 reverse Treg-mediated suppression (51) and that IL-6 together with TGFβ induced pathogenic Th17 cells (52). In initial experiments, we have found that CD4⁺CD25⁻LAP⁺ cells could be induced to differentiate into CD4⁺CD25⁺LAP⁺ cells in the presence of TGFβ (M.-L. Chen and H. L. Weiner, unpublished observations). Nonetheless, the exact relationship between CD4⁺CD25⁺LAP⁺, CD4⁺CD25⁻LAP⁺ and the Th3 Tregs that appear after oral administration of Ag (20) remains to be determined.

In summary, we have characterized a novel subset of CD4⁺CD25⁺ Tregs that express LAP on their surface, and these cells possess enhanced suppressive properties. Our findings further define the characteristics and properties of CD4⁺CD25⁺ Tregs. Induction of Tregs is one of the major goals for the immunotherapy of autoimmune diseases. We have previously shown that oral administration of CD3-specific Ab is able to induce or expand CD4⁺CD25⁻LAP⁺ and CD4⁺CD25⁺LAP⁺ populations (26). Thus, it may be possible to target CD4⁺CD25⁺LAP⁺ cells for therapeutic purposes.