Regulatory T cells (Tregs) are found at sites of chronic inflammation where they mediate bystander and Ag-specific suppression of local immune responses. However, little is known about the molecular control of Treg recruitment into inflamed human tissues. We report that up to 18% of T cells in areas of inflammation in human liver disease are forkhead family transcriptional regulator box P3 (FoxP3)+ Tregs. We isolated CD4+CD25+CD127lowFoxP3+ Tregs from chronically inflamed human liver removed at transplantation; compared with blood-derived Tregs, liver-derived Tregs express high levels of the chemokine receptors CXCR3 and CCR4. In flow-based adhesion assays using human hepatic sinusoidal endothelium, Tregs used CXCR3 and α4β1 to bind and transmigrate, whereas CCR4 played no role. The CCR4 ligands CCL17 and CCL22 were absent from healthy liver, but they were detected in chronically inflamed liver where their expression was restricted to dendritic cells (DCs) within inflammatory infiltrates. These DCs were closely associated with CD8 T cells and CCR4+ Tregs in the parenchyma and septal areas. Ex vivo, liver-derived Tregs migrated to CCR4 ligands secreted by intrahepatic DCs. We propose that CXCR3 mediates the recruitment of Tregs via hepatic sinusoidal endothelium and that CCR4 ligands secreted by DCs recruit Tregs to sites of inflammation in patients with chronic hepatitis. Thus, different chemokine receptors play distinct roles in the recruitment and positioning of Tregs at sites of hepatitis in chronic liver disease.

Immune homeostasis is dependent on the ability of the immune system to respond to pathogens while suppressing responses to self-Ags and harmless environmental Ags. The maintenance of tolerance to self-Ags involves central and peripheral mechanisms. T regulatory cells (Tregs) are a subset of T lymphocytes that play a crucial role in maintaining peripheral tolerance by suppressing self-reactive T cells and preventing collateral autoimmune damage as a consequence of immune responses to infections (1). Tregs are also implicated in the tolerogenic properties of the liver and help to suppress responses to harmless food Ags and bacterial components to which the liver is constantly exposed via the portal vein (2).

Tregs have been implicated in several inflammatory liver diseases. It has been suggested that their frequency is increased in the liver in chronic viral hepatitis, in which a failed immune response leads to viral persistence, and in hepatocellular carcinoma, in which protective antitumor immune responses may be suppressed (3, 4). In contrast, there is evidence for reduced Treg frequency and defective function in autoimmune liver disease (5, 6).

Naturally occurring Tregs are thymus-derived cells that are present from birth. They are defined by the expression of CD4, CD25, and the forkhead family transcriptional regulator box P3 (FoxP3), which is required for their development and function (7), and low expression of the IL-7 receptor (CD127) (810). Tregs suppress effector cell proliferation and cytokine production in vitro in a cell contact-dependent manner and protect against autoimmunity following adoptive transfer in vivo (11). In the healthy organism, Tregs suppress effector T cell activation in secondary lymphoid tissues to maintain steady-state self-tolerance. However, in chronic inflammatory states and autoimmune disease, Tregs are detected at sites of inflammation where they suppress immune responses locally (1217). Thus, when steady-state self-tolerance is breached as a consequence of infection or chronic inflammation, Tregs become further activated to allow their migration into peripheral tissue where they limit local bystander and collateral damage. The ability to enter peripheral tissues, such as the liver, requires the downregulation of chemokine receptors and adhesion molecules, such as CCR7 and l-selectin, which promotes entry to secondary lymphoid tissues and the expression of receptors required for recruitment into sites of inflammation. Little is known about the receptors involved in the recruitment and positioning of Tregs in chronically inflamed human tissues. In the current study, we report that chronic inflammatory liver disease is associated with a high frequency of liver-infiltrating CD4+CD25+CD127lowFoxP3+ Tregs that are detected in close association with intrahepatic dendritic cells (DCs) at sites of inflammation. We show that liver-infiltrating Tregs (LITregs) use CXCR3 to undergo transendothelial migration across hepatic sinusoidal endothelium and use CCR4 to respond to CCL17 and CCL22 secreted by intrahepatic DCs. Thus, different chemokine receptors play distinct roles in the recruitment and positioning of Tregs at sites of hepatitis in chronic liver disease.

Venous blood was obtained from patients undergoing liver transplantation before induction of anesthesia and before any blood transfusion. Subsequently, explanted diseased liver was obtained from the same patients to allow the isolation of matched blood- and liver-derived cells. Nondiseased liver tissue was obtained from donor liver tissue surplus to clinical requirements or from uninvolved liver tissue removed at the time of resection for colorectal hepatic metastases. All samples were collected with appropriate patient consent and local research ethics committee approval.

The following primary Abs were used for immunohistochemistry, immunofluorescence, and confocal microscopy: mouse anti-human CCL22 mAb (MAB 336, R&D Systems, Minneapolis, MN), goat anti-human CCL17 mAb (AF 364, R&D Systems), mouse anti-human FoxP3 (Abcam, 236A/E7, Cambridge, U.K.), rabbit anti-human polyclonal to CCL22 amino terminal end (Ab 53002-100; Abcam), rabbit anti-human monoclonal to CD11c (Abcam, EP13474), rabbit monoclonal to phosphorylated STAT5 (phosphoSTAT5)(Abcam/E208), mouse anti-human CD4 IgG 2b (eBioscience, Hatfield, U.K.), anti-human CD3 IgG2a (eBioscience), and mouse anti-human CD8 IgG2a (eBioscience). Relevant secondary Abs used were anti-IgG1 FITC (Southern Biotech, Cambridge, U.K.), rabbit-anti FITC (Sigma-Aldrich, St. Louis, MO), goat-anti rabbit FITC (Southern Biotech), anti-IgG2b biotinylated (Southern Biotech), anti-IgG1 biotinylated (Southern Biotech), streptavidin 555 (Invitrogen, Paisley, U.K.), streptavidin 647 (Invitrogen), mouse anti-human IgG2a Cy-3 (Southern Biotech), and goat anti-mouse IgG1 FITC (Southern Biotech).

The following Abs were used for multicolor flow cytometry: mouse anti-human CD4 Pacific Blue (RPA-T4, BD Pharmingen, San Diego, CA), mouse anti-human CD25 PE Cy-5 (IM2646, Beckman Coulter, Fullerton, CA), mouse anti-human CD3 PE-Cy7 (UCHT-1, eBioscience), mouse anti-human CD127 Alexa-488 (Biolegend, Cambridge, U.K.; HCD127), mouse anti-human APC Cy-7 CD27 (O323, Biolegend), mouse anti-human CD39 PE (Clone eBio A1, eBioscience), mouse anti-human CD134/OX40 (555838, BD Pharmingen), rat anti-human Foxp3 PE (PCH 101, eBioscience), and CD45 RO PE Cy-5 (UCHL-1, BD Pharmingen). CCR1 mAb (53504/ FAB145P, 1/10), CCR2 mAb (48607/FAB151P), CCR3 mAb (61828/FAB155P), CCR4 mAb (205410/FAB 1567P, FAB1567F), CCR5 mAb (CTC5/FAB1802P), CCR6 mAb (53103/FAB 195P), CCR7 mAb (150503/FAB197P), CCR8 mAb (191704/FAB1429P), CCR9 mAb (112509/FAB179P), CCR10 mAb (FAB3478P), CXCR1 mAb (42705/FAB330P), CXCR2 mAb (48311/FAB331P), CXCR3 mAb (FAB160P, FAB160F), CXCR3 mAb (550967; BD), CXCR4 mAb (12G5/FAB170P), CXCR5 mAb (51505/FAB190P), and CXCR6 mAb (56811/FAB699P) were purchased from R&D Systems. For intracellular cytokine staining, FITC-conjugated mAb to IFN-γ (340449, BD Pharmingen), FITC-conjugated mAb to IFN-γ (MHCIFG05, Caltag Laboratories, Buckingham, U.K.), and FITC-conjugated mAb to TNF-α (554512, BD Pharmingen) were used.

Liver-infiltrating lymphocytes (LILs) were isolated from fresh liver tissue, as previously described (18, 19). To maximize yield and viability, LITregs were isolated immediately after explantation of livers from patients with various liver diseases, including alcoholic cirrhosis, nonalcoholic steatohepatitis, primary biliary cirrhosis (PBC), and autoimmune hepatitis (AIH). Liver tissue was diced into 5-mm3 cubes and subjected to mechanical homogenization in a Seward stomacher 400 circulator (230 rpm for 5 min), without the use of an enzymatic-digestion step, which removes some chemokine receptors. The resulting homogenate was filtered through fine gauze, and lymphocytes were separated using Lympholyte (VH Bio, Gateshead, U.K.) density gradient centrifugation (650 × g for 30 min). The lymphocyte band was removed and washed, and cell viability was assessed by trypan blue exclusion. Tregs were purified from the total lymphocyte population by negative selection to remove CD8+ cells and positive selection of CD4+CD25+ cells with a Treg isolation kit (DYNAL, Invitrogen). LITregs were defined using cell surface staining as CD3+, CD4+, CD25high, or CD127low, and intracellular staining (fix & perm kit from eBioscience) was used to confirm the expression of FoxP3.

The purity of Tregs isolated using the CD4+CD25+ Treg kit was confirmed to be >95% by flow cytometry, and Treg status was confirmed by Foxp3 RT-PCR (data not shown). Following isolation, Tregs were cultured in RPMI 1640, 10% human serum, streptomycin (100 U/ml), penicillin (100 IU/ml), glutamine (2 mmol/l), and IL-2 (500 U/ml) in round-bottom 96-well culture plates.

Lymphocytes were isolated from peripheral venous blood by density gradient centrifugation over a Lympholyte gradient (30 min at 650 × g). Peripheral blood Tregs (PBTregs) were purified, as described above, using a Treg isolation kit and phenotyped using CD4, CD25high, CD127low, and intracellular FoxP3 expression to define Tregs. Purity was confirmed by flow cytometry, as described above.

Liver-infiltrating DCs (LIDCs) were isolated from diseased liver using a method established in our laboratory (20). Briefly, after isolation of lymphocytes, followed by a 13.5% OptiPrep density gradient (Axis Shield, Cambridge, U.K.), CD11c-positive magnetic selection was used to obtain highly purified DCs. Monocyte-derived DCs (MoDCs) from peripheral blood were isolated according to established methods, as previously described (21). Briefly, OptiPrep was added to buffy coats made from whole blood to increase its density to ∼1.1 g/ml, overlayered by two different density solutions of OptiPrep (1.068 and 1.084 g/ml), and centrifuged at 750 × g for 20 min. The monocytes in the 1.068-g/ml layer were collected and cultured for 5 d in media containing RPMI 1640, 5% FCS, and 10 ng/ml IL-4 (BD Pharmingen) with 1000 U/ml recombinant human GM-CSF (Mielogen; Schering-Plough, Kenilworth, NJ), after which the immature DCs were matured with TNF-α. To obtain highly pure DCs, these cells underwent positive immunomagnetic selection with CD11c.

Hepatic sinusoidal endothelial cells (HSECs) were isolated according to methods established in our laboratory (22). Explanted liver tissue (∼30 g) was finely chopped and subjected to enzymatic digestion (1 mg/ml, collagenase type IV; Sigma-Aldrich) and density gradient centrifugation (Percoll; Amersham Biosciences, GE Healthcare, Little Chalfont, U.K.). The nonparenchymal cell band was removed and further purified by immunomagnetic selection using mAb HEA-125 (PROGEN, Heidelberg, Germany) to remove cholangiocytes. CD31 (DAKO) was used to select HSECs, which were plated in collagen-coated 25-cm2 tissue culture flasks (Corning Glass, Londonborough, U.K.) and cultured in human endothelial basal media (Invitrogen) containing 10% human serum (H&D Supplies, Bucks, U.K.), penicillin, streptomycin (100 IU/ml), glutamine (2 mmol/l), hepatocyte growth factor (10 ng/ml, PeproTech, Peterborough, U.K.), and vascular endothelial growth factor (10 ng/ml, PeproTech). Cells were cultured to confluence and were used between passages two and four in all experiments.

Formalin-fixed paraffin-embedded liver specimens from seven cases each of normal liver (NL), AIH, alcoholic liver disease (ALD), PBC, fulminant seronegative hepatitis or non-A non-B hepatitis, chronic hepatitis B virus, and chronic hepatitis C virus (HCV); five portal draining lymph nodes were examined. The tonsil/lymph node tissues were used as positive control, and negative controls were used for each staining as well. The liver tissue sections were dewaxed in xylene for 10 min and incubated through different strengths of alcohol. Endogenous peroxidase was blocked for 20 min (1% hydrogen peroxide in methanol), and Ag retrieval was achieved using EDTA buffer. Then slides were incubated in horse serum for 20 min, followed by incubation in primary Abs CD3 (clone F7.2.38; DakoCytomation) and FoxP3 (clone 236A/E7; Abcam) and rabbit mAb to phosphoSTAT5 (Y694; Abcam) in blocking buffer (2% horse serum) for 60 min at 37°C in a humidified container for 1 h. The slides were washed in TBS/Tween (pH 7.6) for 5 min and incubated in Vector ImmPRESS Universal secondary kit (mouse anti-human and rabbit anti-human, Vector Laboratories, U.K.) for 30 min and diaminobenzidine substrate to visualize staining. Slides were counterstained in hematoxylin and mounted in DPX.

For double immunohistochemical staining, formalin-fixed paraffin-embedded tissue sections were dewaxed, rehydrated in TBS/Tween and blocked with endogenous peroxidase (as described above). Sections were subjected to Ag retrieval and incubated with 2% casein (Vector Laboratories) for 10 min. Primary mouse mAb, FoxP3 (clone 236A/E7, Abcam) was applied at a 1/50 dilution for 30 min. After a TBS/Tween wash, secondary ImmPRESS anti-mouse reagent (Vector Laboratories) was applied and visualized with ImmPACT diaminobenzidine (Vector Laboratories) for 5 min. Then sections were incubated again with 2% casein, followed by primary rabbit anti-human polyclonal CCL22 (Ab 53002, Abcam); secondary ImmPRESS rabbit anti-human reagent was applied, and visualization was performed with NovaRed substrate (both from Vector Laboratories). Sections were counterstained in hematoxylin, dehydrated through alcohol, placed into xylene, and mounted in DPX.

Sections for dual immunofluorescence were incubated with 10% normal mouse or rabbit serum for 30 min before adding the primary Abs against CCL22 or CD11c for 1 h. Control sections were incubated with IgG mouse mAb or rabbit IgG. Subsequently, sections were incubated with FITC or Texas Red secondary Abs for 30 min in the dark. Nuclei were counterstained with DAPI. Dual immunofluorescence was assessed using Axio Vision software.

Liver tissue sections for NLs, diseased livers, and portal lymph nodes were incubated with blocking buffer (10% horse serum in PBS/1% BSA). Primary Abs against FoxP3, CD11c, CD3, and CD4 were applied for 1 h. FoxP3 primary Abs were amplified with anti-IgG1 biotin, followed by streptavidin Alexa Fluor 555 (Invitrogen). Rabbit monoclonal CD11c was visualized by goat anti-rabbit FITC, then with rabbit anti-FITC, and, finally, with goat anti-rabbit FITC. CD8 and CD4 Abs were detected with anti-IgG2a Cy-3/IgG2b tetramethylrhodamine isothiocyanate (TRITC). All incubation steps were for 30 min, with 10-min washes in PBS, in a humidified chamber protected from the light. Quadruple immunofluorescence staining was visualized and analyzed by confocal microscopy using LSM software (ZEISS).

DCs and HSECs were stimulated with cytokines for 24 h, and cells and supernatant were stored for later analysis by RT-PCR and ELISA.

LIDCs and MoDCs were stored in RNA later, and CCL17 and CCL22 gene expression was analyzed by RT-PCR amplification of RNA from LPS-stimulated and unstimulated LIDCs and MoDCs and from TNF-α + IFN-γ–stimulated and unstimulated HSEC low passage (passage two through four) and high passage (passage five and above).

Total RNA was extracted from cells using an RNeasy kit (Qiagen, Crawley, U.K.), and single-stranded cDNA was synthesized. The PCR reaction was conducted using hot start 95°C for 5 min, denaturing at 95°C for 3 min, annealing at 66°C, extension at 72°C for 33 cycles, and final extension at 72°C. Samples were stored at 4°C until used. Primers were designed from GenBank sequences for GAPDH, CCL17, and CCL22: GAPDH forward, 5′-GCC AAG GTC ATC CAT GAC AAC TTT GG-3′, reverse, 5′-GCC TGC TTC ACC ACC TTC TTG ATG TC-3′; CCL17 forward, 5′-ACT GCT CCA GGG ATG CCA TCG TTT TT-3′, reverse, 5′-ACA AGG GGA TGG GAT CTC CCT CAC TG-3′; and CCL22 forward, 5′-AGG ACA GAG CAT GGA TCG CCT ACA GA-3′, reverse, 5′-TAA TGG CAG GGA GGT AGG GCT CCT GA-3′. Product sizes for each set of primers were 270 bp for CCL17 and 363 bp for CCL22. Positive (MoDCs) and negative controls (no cDNA) were included in each assay. Amplified products were analyzed on a 2% agarose gel containing ethidium bromide (Sigma-Aldrich).

Chemokines in cell culture supernatants (normalized to cell numbers) were measured by sandwich ELISA using Quantikine Human Immunoassay kits specific for human CCL17 (DDN00) and CCL22 (DMO00), according to the manufacturer’s instructions (both from R&D Systems). All measurements were performed using triplicate samples for each experiment.

We used freshly isolated LILs and matched PBLs for phenotypic analysis. Cells were washed, resuspended, and labeled for 45 min with different fluorochrome-labeled primary Abs against Treg surface markers, phenotypic markers, and chemokine receptors at optimal dilutions at 4°C before washing with PBS 5% FCS. Following surface staining, cells were fixed in paraformaldehyde and permeabilized using a permeabilization buffer containing saponin (eBioscience, Cat00-5523), FoxP3 Ab conjugated to PE was used to detect intracellular FoxP3, and 2% rat serum was used to reduce background staining. Control samples were labeled with matched isotype control Abs. PBLs and LILs were gated using forward and side scatter parameters and then regated on CD3 and CD4. CD25high, CD127low, and FoxP3+ were used to define Tregs and chemokine receptor expression, and immunophenotyping was done on this population.

To detect intracellular cytokines IFN-γ and TNF-α, LILs were stimulated again with 500 ng/ml ionomycin (Sigma-Aldrich) and/or 50 ng/ml PMA for 5 h in a 37°C incubator. The intracellular cytokine secretion was prevented using a Golgi block (brefeldin A, 1000 ng/ml). Cells were washed, and Abs were added in 100 μl permeabilization buffer. Samples were analyzed on a nine-color Dako Cyan Flow Cytometer using Summit 4.3 software (DakoCytomation).

Liver-infiltrating CD4 cells were isolated by negative selection and divided into CD25+CD4+ and CD25CD4+ subsets by magnetic cell sorting. The ability of CD25+CD4+ cells (>95% purity by flow cytometry) to suppress proliferation was assessed by culturing liver-derived CD25+CD4+ lymphocytes with autologous naive lymphocytes, at ratios of 1:1 to 1:16, with added allogenic irradiated liver-derived DCs (1:100 ratio) or anti-human CD3/CD28 beads (DYNAL, Invitrogen) in 96-well round-bottom plates (Nunc, Rochester, NY). Suppression of proliferation was assessed by the incorporation of [3H]thymidine. Cultures were pulsed with 1 μCi per well [3H]thymidine (Amersham Biosciences) for the last 16 h of a 5-d incubation at 37°C in 5% CO2, the plates were harvested, and proliferation was assessed by measuring radioactivity with a liquid scintillation counter. All cells were cultured in triplicate in a final volume of 200 μl RPMI 1640 with l-glutamine (2 mmol/l), 100 IU/ml penicillin, 100 IU/ml streptomycin, and 10% human AB serum.

The ability of LITregs to migrate to chemokines was assessed using 24-well plate transwells and fibronectin-covered (Sigma-Aldrich), 6.5-mm- diameter, 5-μm pore inserts (Corning). The lower chambers were filled with 600 μl assay media (RPMI 1640 medium with 0.1% BSA; Sigma-Aldrich) alone or supplemented with chemokines (100 ng/ml recombinant human CCL22 and CCL17, both from R&D Systems). Freshly isolated LITregs (5 × 105 in RPMI 1640) were added to the upper chamber. LITregs were collected from the top and bottom chambers after 4 h incubation at 37°C in 5% CO2; the proportion of transmigrated cells was calculated by fixed-volume counting and phenotyped for Treg markers and chemokine receptor expression by flow cytometry. Assays were conducted in duplicate and compared with control wells containing RPMI 1640 medium with 0.1% BSA alone. Tregs were incubated with pertussis toxin (PTx; 100 ng/ml; Sigma-Aldrich) prior to the assay to inhibit chemokine-mediated signaling.

We performed a flow-based adhesion assay incorporating HSECs grown in microslides to study interactions between human HSECs and freshly isolated Tregs under physiological blood flow (18, 22). HSECs were cultured to confluence in glass microslides, stimulated for 24 h with TNF-α and IFN-γ (both 10 ng/ml; PeproTech), and connected to the flow-based system, as described previously (22). Freshly isolated Tregs were perfused through microslides at flow rates of 0.182 ml/min, generating shear stresses of 0.05 Pa, which are comparable to wall shear stresses found in the liver sinusoids in vivo. Adherent cells were visualized microscopically under phase contrast, and a video monitor was used to record assays for off-line analysis to determine the percentage of rolling cells, statically adherent cells, and transmigrated cells. The number of adherent cells was converted to adherent cells/mm2 corrected for the number of Tregs perfused (i.e., adherent cells/mm2 per 106 Tregs perfused). Cells appearing phase bright were above the endothelial monolayer, whereas those appearing phase dark had migrated through the monolayer. HSEC monolayers were incubated with blocking Abs against ICAM-1 (11C81: 10 μg/ml) and VCAM-1 (4B2: 10 μg/ml) (both from R&D Systems), or lymphocytes were incubated with a blocking Ab raised against CXCR3 (MAB160/49801; 10 μg/ml; R&D Systems;) for 30 min. To inhibit chemokine-mediated signaling, Tregs were incubated with PTx (100 ng/ml; Sigma-Aldrich) 30 min prior to assay. We also blocked the individual CXCR3 chemokine ligands by incubating HSECs with anti-CXCL9 (49106; 50 μg/ml), anti-CXCL10 (33036; 5 μg/ml), or anti-CXCL11 (87328; 5 μg/ml) (all from R&D Systems) for 30 min. HSECs had been stimulated for 24 h with TNF-α and IFN-γ to induce the expression of CXCR3 ligands.

All experiments were compared with control microslides in which isotype-matched IgG was used.

Paired or independent t tests were used to assess data that were normally distributed, whereas nonparametric data were compared using the Mann-Whitney U test.

Liver and blood-derived Tregs were defined by the characteristic phenotype CD4+CD25highCD127lowFoxP3+ (8, 9). Liver-derived lymphocytes were freshly isolated from explanted human liver tissue by mechanical digestion and density gradient centrifugation before being stained with Abs against CD3, CD4, CD25, CD127, and the intracellular transcription factor FoxP3. More than 95% of CD4+CD25highCD127low LILs expressed FoxP3 (Fig. 1A). The vast majority of LITregs expressed CD45RO, and >80% expressed high levels of CD27 (Fig. 1B), consistent with a tissue-infiltrating phenotype. More than 90% of LITregs expressed the surface ectoenzyme CD39, and 35% expressed OX-40 (Fig. 1B).

FIGURE 1.

Phenotype of LITregs freshly isolated from explanted human liver without expansion in vitro. A, Human LILs were stained for cell surface expression of CD4, CD25, and CD127. The stained cells were fixed, permeabilized, and stained intracellularly for FoxP3. For analysis, LILs were gated on forward scatter and side scatter, and LITregs were identified by gating on CD3highCD4high lymphocytes, followed by CD25high (y-axis) and CD127low (x-axis). The square box represents the population of LITregs. Graph shows the expression of transcription factor FoxP3, which was analyzed on LITregs (CD3highCD4highCD25highCD127low) by flow cytometry. B, Flow cytometry overlay histograms show the percentage of LITregs positive for CD45RO, CD27, CD39, or OX-40. The percentage expression of relevant markers are provided. Representative data of one of six independent experiments. C, The suppressive function of freshly isolated human LITregs was assessed by [3H]thymidine proliferation. LITregs were isolated using the Dynal Treg isolation kit from freshly isolated liver-derived lymphocytes. Suppression of proliferation was done at ratios of 1:16 to 1:1 freshly isolated LITregs/naive lymphocytes activated by allogenic liver-derived DCs. Data represent mean ± SEM of three independent experiments. *p < 0.05; **p < 0.005 by Student t test.

FIGURE 1.

Phenotype of LITregs freshly isolated from explanted human liver without expansion in vitro. A, Human LILs were stained for cell surface expression of CD4, CD25, and CD127. The stained cells were fixed, permeabilized, and stained intracellularly for FoxP3. For analysis, LILs were gated on forward scatter and side scatter, and LITregs were identified by gating on CD3highCD4high lymphocytes, followed by CD25high (y-axis) and CD127low (x-axis). The square box represents the population of LITregs. Graph shows the expression of transcription factor FoxP3, which was analyzed on LITregs (CD3highCD4highCD25highCD127low) by flow cytometry. B, Flow cytometry overlay histograms show the percentage of LITregs positive for CD45RO, CD27, CD39, or OX-40. The percentage expression of relevant markers are provided. Representative data of one of six independent experiments. C, The suppressive function of freshly isolated human LITregs was assessed by [3H]thymidine proliferation. LITregs were isolated using the Dynal Treg isolation kit from freshly isolated liver-derived lymphocytes. Suppression of proliferation was done at ratios of 1:16 to 1:1 freshly isolated LITregs/naive lymphocytes activated by allogenic liver-derived DCs. Data represent mean ± SEM of three independent experiments. *p < 0.05; **p < 0.005 by Student t test.

Close modal

Allogenic stimulation/suppression assays confirmed that LITregs suppress T cell activation in vitro at ratios of 1:4 Treg/responder cells, demonstrating that LITregs are functional (Fig. 1C).

Few Tregs were detected in NL tissue by immunohistochemistry, although they could be readily detected in the chronically inflamed livers (Fig. 2A) in septal and parenchyma areas (Fig. 2A, 2B). Multicolor confocal microscopy staining revealed them to be closely associated with CD8 T cells and CD11c DCs (Fig. 2B). Tregs were also detected in liver-draining portal lymph nodes in close association with CD8+ cells and CD11c+ DCs (Fig. 2C). T cells were readily isolated from normal and inflamed liver tissue (Fig. 2D), and the total numbers and the proportion of the infiltrate that was CD4+CD25highCD127lowFoxP3+ was greater in diseased tissue than in NL tissue (Fig. 2D). We used immunohistochemistry to quantify the proportion of the CD3 infiltrate that was FoxP3+ in a variety of liver diseases. In all diseases studied, the inflamed diseased liver contained a higher proportion of Tregs compared with nondiseased liver (Fig. 2E). Interestingly, greater numbers of Tregs were seen in the patients with the most severe inflammation, with the highest frequencies detected in patients who had undergone emergency liver transplantation for fulminant liver failure as a consequence of severe seronegative (non-A non-B) hepatitis (Fig. 2E). In seronegative hepatitis and AIH, FoxP3+ Tregs were found at higher frequencies in the parenchyma compared with PTs, whereas in the other liver diseases and in NL, FoxP3+ Tregs were detected at higher frequencies in PTs (Fig. 2F)

FIGURE 2.

Detection of Tregs in different liver diseases and portal lymph node. Ai, Single immunohistochemical staining of formalin-fixed, paraffin-embedded liver sections revealed FoxP3+ cells staining using mAb (Abcam clone 236A/E7) (detected as a brown pigment) in PT and parenchyma (P) of patients with NL and different chronic liver diseases: AIH, ALD, non-A non-B (seronegative) hepatitis, PBC, chronic hepatitis B (HBV), and chronic HCV. FoxP3+ cells in liver tissue are indicated by arrows. Control sections are stained with isotype-matched IgG (original magnification ×100). Aii, Expression of FoxP3+, CD3+, and CD4+ cells in inflamed liver (AIH and PBC shown as representative examples). Right panel shows magnification of left panel; arrow indicates costaining of Foxp3, CD3, and CD4. B, Multicolor confocal microscopy images of chronically inflamed liver compared with NL stained with mouse anti-human mAb to FoxP3 IgG1 (red), rabbit anti-human mAb to CD11c (green), and mouse anti-human mAb to CD8 IgG2b (blue). Cell nuclei stained with DAPI (gray). Staining was detected in PT, septa (i), and parenchyma (ii). Top panel, original magnification ×10 (scale bar, 200 μ); bottom panel original magnification ×40 (scale bar, 50 μ).C, Confocal images of CD8+ (blue), FoxP3 (red), and CD11c (green) staining in a portal lymph node from a patient with PBC (left panel; representative staining from four experiments). The right panel (i) shows the magnified view (×40) of the boxed area in the left panel (×10). D, Flow cytometry dot-plot comparing the frequency of LITregs in normal and inflamed diseased liver (representative plots shown from eight diseased livers and seven NLs). LITregs were identified by gating on LILs CD3highCD4high, followed by CD25high (y-axis) and CD127low (x-axis). E, The ratio of FoxP3+ cells to CD3-staining T cells was quantified by immunohistochemistry in NL and different chronic liver diseases. p < 0.05 comparing normal versus diseased livers; Mann-Whitney U test (n = 8). F, Percentage of FoxP3+ cells in relation to CD3 in PT and parenchyma are shown for NL and different chronic liver diseases (n = 7).

FIGURE 2.

Detection of Tregs in different liver diseases and portal lymph node. Ai, Single immunohistochemical staining of formalin-fixed, paraffin-embedded liver sections revealed FoxP3+ cells staining using mAb (Abcam clone 236A/E7) (detected as a brown pigment) in PT and parenchyma (P) of patients with NL and different chronic liver diseases: AIH, ALD, non-A non-B (seronegative) hepatitis, PBC, chronic hepatitis B (HBV), and chronic HCV. FoxP3+ cells in liver tissue are indicated by arrows. Control sections are stained with isotype-matched IgG (original magnification ×100). Aii, Expression of FoxP3+, CD3+, and CD4+ cells in inflamed liver (AIH and PBC shown as representative examples). Right panel shows magnification of left panel; arrow indicates costaining of Foxp3, CD3, and CD4. B, Multicolor confocal microscopy images of chronically inflamed liver compared with NL stained with mouse anti-human mAb to FoxP3 IgG1 (red), rabbit anti-human mAb to CD11c (green), and mouse anti-human mAb to CD8 IgG2b (blue). Cell nuclei stained with DAPI (gray). Staining was detected in PT, septa (i), and parenchyma (ii). Top panel, original magnification ×10 (scale bar, 200 μ); bottom panel original magnification ×40 (scale bar, 50 μ).C, Confocal images of CD8+ (blue), FoxP3 (red), and CD11c (green) staining in a portal lymph node from a patient with PBC (left panel; representative staining from four experiments). The right panel (i) shows the magnified view (×40) of the boxed area in the left panel (×10). D, Flow cytometry dot-plot comparing the frequency of LITregs in normal and inflamed diseased liver (representative plots shown from eight diseased livers and seven NLs). LITregs were identified by gating on LILs CD3highCD4high, followed by CD25high (y-axis) and CD127low (x-axis). E, The ratio of FoxP3+ cells to CD3-staining T cells was quantified by immunohistochemistry in NL and different chronic liver diseases. p < 0.05 comparing normal versus diseased livers; Mann-Whitney U test (n = 8). F, Percentage of FoxP3+ cells in relation to CD3 in PT and parenchyma are shown for NL and different chronic liver diseases (n = 7).

Close modal

The results of staining for chemokine receptor expression on freshly isolated liver-derived Tregs are shown in Fig. 3A. Approximately 20% ± 5% of LITregs expressed CCR7, but the majority expressed CCR4 (74% ± 7%) and CXCR3 (67 ± 9% SD) (Fig. 3A). Furthermore, the levels of these receptors were higher on LITregs isolated from inflamed liver compared with NL tissue: 74% ± 7%% versus 32% ± 7% for CCR4 and 67% ± 9% versus 21% ± 6% for CXCR3 (p ≤ 0.05 for both) (Fig. 3C). We compared levels of CCR4 and CXCR3 in circulating and liver-derived Tregs of patients from whom we had collected matched blood and liver samples (Fig. 3B): 82% of LITregs stained for CCR4 versus 67% of matched blood Tregs, and 66% of LITregs stained for CXCR3 compared with 43% of matched blood Tregs (p < 0.05). Seventy-four percent of LITregs expressed CCR4, compared with 40% of liver-infiltrating CD4+CD25low cells from diseased livers. Thus, CCR4 is preferentially expressed on LITregs. However, the levels of CXCR3 were similar for LITregs and liver-infiltrating CD4+CD25low cells (Fig. 3C).

FIGURE 3.

Chemokine receptor expression on LITregs. A, Flow cytometry was used to analyze staining for chemokine receptors on liver-infiltrating CD4highCD25highCD127lowFoxP3+ Tregs freshly isolated from diseased liver tissues. The chemokine receptors CCR1-10 and CXCR1-6 were stained together with Treg markers, and the percentage of cells expressing each chemokine receptor was analyzed. Data represent mean ± SEM of seven different experiments using livers from patients with ALD, PBC, cryptogenic cirrhosis, AIH, seronegative (non-A non-B) hepatitis, and chronic hepatitis C. B, Overlay graph of CCR4 and CXCR3 frequency on LITregs from inflamed liver tissue compared with peripheral blood Tregs (PBTregs) in matched samples from patients who underwent liver transplantation. Cells were gated on CD4highCD25highCD127lowFoxP3+. One representative graph of five experiments from a patient with PBC. C, Analysis of CXCR3 expression on freshly isolated CD4highCD25highLITregs compared with CD4highCD25low cells from diseased livers (top panel), CXCR3 expression on diseased CD4highCD25highLITregs compared with normal CD4highCD25high LITreg (middle panel), and CCR4 expression on CD4highCD25high LITregs compared with CD4highCD25low cells from diseased liver (bottom panel). LILs were gated on Treg markers, CD4, CD25, and CD4CD25high; CD4CD25low gating was applied to CXCR3 and CCR4 expression. Data represent mean ± SEM (n = 5). *p < 0.05, Student t test. D, CCR4 and CXCR3 are coexpressed on LITregs. LILs gated on CD3+CD4+CD127lowCD25high were stained for CCR4 (y-axis) and CXCR3 (x-axis), confirming the coexpression of these chemokine receptors on the same cell. One representative experiment of five is shown from a patient with AIH.

FIGURE 3.

Chemokine receptor expression on LITregs. A, Flow cytometry was used to analyze staining for chemokine receptors on liver-infiltrating CD4highCD25highCD127lowFoxP3+ Tregs freshly isolated from diseased liver tissues. The chemokine receptors CCR1-10 and CXCR1-6 were stained together with Treg markers, and the percentage of cells expressing each chemokine receptor was analyzed. Data represent mean ± SEM of seven different experiments using livers from patients with ALD, PBC, cryptogenic cirrhosis, AIH, seronegative (non-A non-B) hepatitis, and chronic hepatitis C. B, Overlay graph of CCR4 and CXCR3 frequency on LITregs from inflamed liver tissue compared with peripheral blood Tregs (PBTregs) in matched samples from patients who underwent liver transplantation. Cells were gated on CD4highCD25highCD127lowFoxP3+. One representative graph of five experiments from a patient with PBC. C, Analysis of CXCR3 expression on freshly isolated CD4highCD25highLITregs compared with CD4highCD25low cells from diseased livers (top panel), CXCR3 expression on diseased CD4highCD25highLITregs compared with normal CD4highCD25high LITreg (middle panel), and CCR4 expression on CD4highCD25high LITregs compared with CD4highCD25low cells from diseased liver (bottom panel). LILs were gated on Treg markers, CD4, CD25, and CD4CD25high; CD4CD25low gating was applied to CXCR3 and CCR4 expression. Data represent mean ± SEM (n = 5). *p < 0.05, Student t test. D, CCR4 and CXCR3 are coexpressed on LITregs. LILs gated on CD3+CD4+CD127lowCD25high were stained for CCR4 (y-axis) and CXCR3 (x-axis), confirming the coexpression of these chemokine receptors on the same cell. One representative experiment of five is shown from a patient with AIH.

Close modal

We correlated CCR4 with CXCR3 expression on LITregs to determine whether the receptors were coexpressed on the same cells or whether there were two subpopulations of LITregs defined by CCR4 and CXCR3 expression (Fig. 3D). More than 75% of LITregs coexpressed CCR4 and CXCR3 (Fig. 3D), and 40% of the residual LITregs that were CCR4+ but CXCR3low coexpressed CCR4 and CCR5 (data not shown).

Cultured HSECs express markers that are characteristic of sinusoidal endothelium in vivo, including CD31, lymphatic vessel endothelial receptor 1, liver/lymph node-specific intercellular adhesion molecule-3-grabbing integrin, and liver and lymph node sinusoidal endothelial cell C-type lectin (2325). When stimulated by TNF-α, HSECs express ICAM-1 and VCAM-1 and support the adhesion of lymphocytes under conditions of flow (22). Furthermore, HSECs secrete the CXCR3 ligands CXCL9, CXCL10, and CXCL11 in response to IFN-γ; this is increased if TNF-α is used in combination with IFN-γ (18). Thus, we used HSECs treated with 10 ng/ml IFN-γ and TNF-α in the flow-based adhesion assays in the current study to allow us to model inflamed hepatic sinusoidal endothelium, which expresses ICAM-1, VCAM-1, and CXCR3 ligands in chronic hepatitis (18).

When flowed over cytokine-stimulated HSECs, some of the Tregs displayed brief rolling/tethering interactions followed by arrest and stable adhesion, whereas others arrested without rolling; this behavior was similar to that reported using unfractionated liver-derived lymphocytes (18). The numbers of cells undergoing stable adhesion from flow could be reduced by blocking CXCR3 and, to a lesser extent, VCAM-1 but not by blocking ICAM-1. Global Gα protein inhibition with PTx reduced adhesion to levels seen with anti-CXCR3 blockade, suggesting that CXCR3 is the dominant chemokine receptor involved (Fig. 4A). The proportion of adherent cells that subsequently underwent transendothelial migration was also significantly reduced by CXCR3 block. Again, PTx had a similar inhibitory effect to CXCR3 blockade.

FIGURE 4.

Treg adhesion and transmigration across activated HSECs is dependent on CXCR3 and VCAM-1. A, Effect of Gi protein inhibition using PTx and Ab inhibition of CXCR3 or VCAM-1 on total adhesion of Tregs. Flow-based adhesion was done with freshly isolated Tregs flowed across 24-h TNF-α– and IFN-γ–stimulated primary human HSECs. Data represent the number of adherent Tregs/mm2/million perfused (mean ± SEM of five experiments). *p < 0.05; **p < 0.005. Statistical significance was calculated using paired t tests, comparing treatment with control. All experiments were performed at 0.05 Pa, which is similar to the shear stress in hepatic sinusoids in vivo. B, Transmigration of adherent Tregs through the same stimulated HSECs. Migrating cells were calculated using frame-by-frame analysis of experimental videos to count the percentage of phase dark (transmigrated) Tregs. All experiments were performed at 0.05 Pa. Data shown as percentage of adherent Tregs that had transmigrated (mean ± SEM of five experiments). *p < 0.05; paired t tests comparing treatment with control. C, Effect of blocking individual CXCR3 ligands on total adhesion of Tregs. To investigate the individual effect of CXCR3 ligands, anti-CXCL9, anti-CXCL10, and anti-CXCL11 were used to block individual CXCR3 ligands on 24-h TNF-α– and IFN-γ–stimulated primary human HSECs. Adhesion was calculated using frame-by-frame analysis of experimental videos to count the percentage of cells adherent to HSECs. Percentage reduction in adherent Tregs to stimulated endothelium with individual chemokine blockade was compared with control. Data represent mean ± SEM of four experiments. All experiments were carried out at 0.05 Pa. D, Effect of blocking individual CXCR3 ligands on transmigration of Tregs. Transmigration was calculated using frame-by-frame analysis of experimental videos to count the percentage of cells that migrated through HSECs (phase dark). The percentage reduction in transmigration of adherent Tregs to stimulated endothelium with individual chemokine blockade was compared with control. Data represent mean ± SEM of four experiments.

FIGURE 4.

Treg adhesion and transmigration across activated HSECs is dependent on CXCR3 and VCAM-1. A, Effect of Gi protein inhibition using PTx and Ab inhibition of CXCR3 or VCAM-1 on total adhesion of Tregs. Flow-based adhesion was done with freshly isolated Tregs flowed across 24-h TNF-α– and IFN-γ–stimulated primary human HSECs. Data represent the number of adherent Tregs/mm2/million perfused (mean ± SEM of five experiments). *p < 0.05; **p < 0.005. Statistical significance was calculated using paired t tests, comparing treatment with control. All experiments were performed at 0.05 Pa, which is similar to the shear stress in hepatic sinusoids in vivo. B, Transmigration of adherent Tregs through the same stimulated HSECs. Migrating cells were calculated using frame-by-frame analysis of experimental videos to count the percentage of phase dark (transmigrated) Tregs. All experiments were performed at 0.05 Pa. Data shown as percentage of adherent Tregs that had transmigrated (mean ± SEM of five experiments). *p < 0.05; paired t tests comparing treatment with control. C, Effect of blocking individual CXCR3 ligands on total adhesion of Tregs. To investigate the individual effect of CXCR3 ligands, anti-CXCL9, anti-CXCL10, and anti-CXCL11 were used to block individual CXCR3 ligands on 24-h TNF-α– and IFN-γ–stimulated primary human HSECs. Adhesion was calculated using frame-by-frame analysis of experimental videos to count the percentage of cells adherent to HSECs. Percentage reduction in adherent Tregs to stimulated endothelium with individual chemokine blockade was compared with control. Data represent mean ± SEM of four experiments. All experiments were carried out at 0.05 Pa. D, Effect of blocking individual CXCR3 ligands on transmigration of Tregs. Transmigration was calculated using frame-by-frame analysis of experimental videos to count the percentage of cells that migrated through HSECs (phase dark). The percentage reduction in transmigration of adherent Tregs to stimulated endothelium with individual chemokine blockade was compared with control. Data represent mean ± SEM of four experiments.

Close modal

These findings suggest that CXCR3 is crucial for Treg adhesion to, and subsequent transmigration across, inflamed hepatic sinusoids (Fig. 4B). To determine which of the CXCR3 ligands is involved, we repeated the experiments using blocking Abs against the individual CXCR3-binding chemokines CXCL9 (monokine induced by IFN-γ), CXCL10 (IFN-γ inducible protein-10), and CXCL11 (IFN-inducible T cell α-chemoattractant). We observed a reduction in adhesion (Fig. 4C) and transmigration (Fig. 4D) with each individual blocking Ab treatment, but these were less than the inhibition seen with anti-CXCR3, which suggests that all three ligands may be involved. This is consistent with our previous work showing that HSECs secrete all three ligands in response to IFN-γ and TNF-α (18).

To determine whether proinflammatory cytokines secreted by liver-infiltrating effector cells are responsible for activating the CXCR3 pathway, we investigated whether the levels of CXCR3 expression on Tregs correlated with IFN-γ and TNF-α secretion by effector cells in the same livers. NL and inflamed liver contained IFN-γ– and TNF-α–secreting CD4 and CD8 T cells (Fig. 5B). The proportion of cytokine-expressing cells was similar in normal and diseased tissue, although the total number of T cells was much higher in the inflamed livers. Once again, we detected higher frequencies of CXCR3+ Tregs in inflamed liver tissue (60–80% cells positive) compared with those isolated from nondiseased liver (15–28%) (Fig. 5A). However, there was no correlation between CXCR3 expression on Tregs and the proportion of CD4 or CD8 TNF-α– and IFN-γ–secreting cells in the infiltrate (Fig. 5A).

FIGURE 5.

LITreg CXCR3 expression and liver-infiltrating CD4 and CD8 effector cell cytokine expression. To determine whether the levels of CXCR3 expression by LITregs are associated with secretion of TNF-α and IFN-γ by liver-infiltrating CD4 and CD8 effector cells, we correlated levels of CXCR3 expression on LITregs with the frequency of IFN-γ– and TNF-α–secreting effector cells in the same liver samples. A, The percentage of liver-derived CD8 (upper panel) and CD4 (lower panel) T cells expressing TNF-α (left panels) and IFN-γ (right panels) is shown on the x-axis, and the frequency of CXCR3+ Tregs (CD4+CD25highCD127low) isolated from the same donor liver and analyzed without expansion is shown on the y-axis. Cells from four NLs and four diseased livers (AIH, PBC, ALD, and non-A non-B [seronegative] fulminant hepatitis) are shown. There was no clear correlation between CXCR3 frequencies in the Treg population and cytokine secretion by effectors. Intracellular cytokine and TNF-α and IFN-γ staining was done by incubating LILs with PMA and ionomycin for 5 h, with brefeldin A Golgi block in the last 2 h, followed by intracellular cytokine staining. The patient with the highest frequency of TNF-α– and IFN-γ–secreting effector cells had fulminant non-A non-B hepatitis. B, A representative flow cytometry dot-plot analysis of T cells isolated from a diseased liver (PBC) is shown to demonstrate IFN-γ and TNF-α secretion. Cells were gated on LILs and gated again on CD3 lymphocytes (one representative example of four experiments).

FIGURE 5.

LITreg CXCR3 expression and liver-infiltrating CD4 and CD8 effector cell cytokine expression. To determine whether the levels of CXCR3 expression by LITregs are associated with secretion of TNF-α and IFN-γ by liver-infiltrating CD4 and CD8 effector cells, we correlated levels of CXCR3 expression on LITregs with the frequency of IFN-γ– and TNF-α–secreting effector cells in the same liver samples. A, The percentage of liver-derived CD8 (upper panel) and CD4 (lower panel) T cells expressing TNF-α (left panels) and IFN-γ (right panels) is shown on the x-axis, and the frequency of CXCR3+ Tregs (CD4+CD25highCD127low) isolated from the same donor liver and analyzed without expansion is shown on the y-axis. Cells from four NLs and four diseased livers (AIH, PBC, ALD, and non-A non-B [seronegative] fulminant hepatitis) are shown. There was no clear correlation between CXCR3 frequencies in the Treg population and cytokine secretion by effectors. Intracellular cytokine and TNF-α and IFN-γ staining was done by incubating LILs with PMA and ionomycin for 5 h, with brefeldin A Golgi block in the last 2 h, followed by intracellular cytokine staining. The patient with the highest frequency of TNF-α– and IFN-γ–secreting effector cells had fulminant non-A non-B hepatitis. B, A representative flow cytometry dot-plot analysis of T cells isolated from a diseased liver (PBC) is shown to demonstrate IFN-γ and TNF-α secretion. Cells were gated on LILs and gated again on CD3 lymphocytes (one representative example of four experiments).

Close modal

We used chemotaxis assays to confirm that CCR4 on LITregs is functional, as demonstrated by the ability of the cells to migrate to the CCR4 ligands CCL17 and CCL22 (Fig. 6A). The chemotactic indices for CCL17 and CCL22 were approximately three times greater than control. We phenotyped the LITregs before chemotaxis assays and confirmed that >75% expressed CCR4.

FIGURE 6.

CCR4 ligands are detected in diseased liver tissue and secreted by LIDCs. A, LITregs migrate toward CCR4 ligands in vitro. Migration of LITregs to CCR4 ligands CCL22 and CCL17 was assessed using fibronectin-coated 5-μm-pore transwell migration chambers. The lower chambers were filled with 600 μl assay media (RPMI 1640 medium with 0.1% BSA) alone or supplemented with chemokines (100 ng/ml recombinant human CCL22 and CCL17). BSA was used as control. A total of 5 × 105 Tregs isolated from inflamed livers (n = 4) were loaded per well in the upper chamber; Tregs were collected from the top and bottom chambers after 4 h of incubation at 37°C in 5% CO2. The proportion of transmigrated cells was calculated by fixed-volume counting and phenotyped for Treg markers and chemokine receptor expression by flow cytometry. To inhibit chemokine-mediated signaling, Tregs were incubated with PTx (100 ng/ml) prior to the assay. Results are expressed as mean ± SEM. LITregs showed significant migration responses to CCL22 and CCL17 compared with migration to BSA, and these were inhibited by PTx. *p < 0.05, Student t test. B, To determine whether liver DCs secrete CCL17 and CCL22, we isolated and cultured myeloid DCs from human liver tissue and stimulated them with LPS (1 μg/ml) for 24 h. CCL17 and CCL22, measured by ELISA, were detected in cell culture supernatants of LPS-stimulated liver-derived DCs isolated from patients with AIH (n = 2) or PBC (n = 1) or from NL (n = 3). All DCs secreted detectable CCL17 and CCL22, with a >2-fold increase in secretion by DCs derived from inflamed liver tissue compared with NL. Data represent mean ± SEM of three experiments using supernatants from different DCs. Values represent the chemokine concentration (pg/ml) of three replicate wells relevant to a standard curve of known concentration. *p ≤ 0.05; **p ≤ 0.01. C, Expression of CCL22 in human liver tissue samples by immunohistochemistry. CCL22 staining (i) was detected on LIDCs within inflammatory infiltrates (positive staining cells brown) in a patient with AIH. No detectable CCL22 was found in NL or when control Abs were used (ii). (Original magnification ×200.) D, Dual staining of FoxP3-expressing cells (brown) and CCL22 (red) in liver tissue from a patient with AIH reveals CCL22+ cells in close association with FoxP3+ Tregs within an inflammatory infiltrate (original magnification ×200). E, Confocal image of dual staining of FoxP3-expressing cells (red/TRITC) and CD11c (green/FITC) in liver tissue from a patient with AIH reveals CD11c+ cells in close association with FoxP3+ Tregs within an inflammatory infiltrate (original magnification ×40). F, Colocalization of DCs and CCL22 in inflamed liver tissue. Serial liver sections from a patient with PBC were stained with CD11c (top panels) and CCL22 (bottom panels). Cells that coexpress CD11c and CCL22 can be determined by comparing staining of CD11c (top right panel) and CCL22 (bottom right panel) at higher power (original magnification ×200). Staining with isotype-matched control Abs was negative (left panels, original magnification ×100.) G, Colocalization of DCs and CCL22 in inflamed liver tissue. Staining of FITC-labeled CCL22 (green) with Texas Red-labeled CD11c (red). Areas of coexpression (top left panel) appear yellow. Nuclei were stained with DAPI (blue). Control Abs demonstrated minimal background tissue fluorescence. The findings are representative of the pattern of immunofluorescence staining seen in tissue sections from four donors each for PBC, ALD, AIH, and HCV (original magnification × 200). H, De novo synthesis of CCR4 ligands CCL17 and CCL22 by liver-derived DCs was confirmed by RT-PCR. LIDCs isolated from liver tissue from patients with AIH and PBC were analyzed unstimulated or after stimulation with LPS. Monocyte-derived TNF-α–activated DCs (MoDCs) are shown as a positive control, and unstimulated and stimulated HSECs were used as a negative control. CCL17 and CCL22 mRNA was detected in LIDCs. Representative data from one of three experiments is shown. I, Western blot analysis of CCL17 in liver tissue lysates from different chronic liver diseases and unstimulated or TNF-α/IFN-γ–stimulated HSEC lysates. CCL17 is present in chronically inflamed diseased livers, but it could not be detected in NL tissue lysates or in lysates of HSECs. One representative blot of five experiments is shown. HSEC Unstim, unstimulated HSECs; HSEC Stim, HSECs stimulated with TNF-α and IFN-γ; NANB, non-A non-B seronegative hepatitis; NASH, nonalcoholic steatohepatitis; PSC, primary sclerosing cholangitis.

FIGURE 6.

CCR4 ligands are detected in diseased liver tissue and secreted by LIDCs. A, LITregs migrate toward CCR4 ligands in vitro. Migration of LITregs to CCR4 ligands CCL22 and CCL17 was assessed using fibronectin-coated 5-μm-pore transwell migration chambers. The lower chambers were filled with 600 μl assay media (RPMI 1640 medium with 0.1% BSA) alone or supplemented with chemokines (100 ng/ml recombinant human CCL22 and CCL17). BSA was used as control. A total of 5 × 105 Tregs isolated from inflamed livers (n = 4) were loaded per well in the upper chamber; Tregs were collected from the top and bottom chambers after 4 h of incubation at 37°C in 5% CO2. The proportion of transmigrated cells was calculated by fixed-volume counting and phenotyped for Treg markers and chemokine receptor expression by flow cytometry. To inhibit chemokine-mediated signaling, Tregs were incubated with PTx (100 ng/ml) prior to the assay. Results are expressed as mean ± SEM. LITregs showed significant migration responses to CCL22 and CCL17 compared with migration to BSA, and these were inhibited by PTx. *p < 0.05, Student t test. B, To determine whether liver DCs secrete CCL17 and CCL22, we isolated and cultured myeloid DCs from human liver tissue and stimulated them with LPS (1 μg/ml) for 24 h. CCL17 and CCL22, measured by ELISA, were detected in cell culture supernatants of LPS-stimulated liver-derived DCs isolated from patients with AIH (n = 2) or PBC (n = 1) or from NL (n = 3). All DCs secreted detectable CCL17 and CCL22, with a >2-fold increase in secretion by DCs derived from inflamed liver tissue compared with NL. Data represent mean ± SEM of three experiments using supernatants from different DCs. Values represent the chemokine concentration (pg/ml) of three replicate wells relevant to a standard curve of known concentration. *p ≤ 0.05; **p ≤ 0.01. C, Expression of CCL22 in human liver tissue samples by immunohistochemistry. CCL22 staining (i) was detected on LIDCs within inflammatory infiltrates (positive staining cells brown) in a patient with AIH. No detectable CCL22 was found in NL or when control Abs were used (ii). (Original magnification ×200.) D, Dual staining of FoxP3-expressing cells (brown) and CCL22 (red) in liver tissue from a patient with AIH reveals CCL22+ cells in close association with FoxP3+ Tregs within an inflammatory infiltrate (original magnification ×200). E, Confocal image of dual staining of FoxP3-expressing cells (red/TRITC) and CD11c (green/FITC) in liver tissue from a patient with AIH reveals CD11c+ cells in close association with FoxP3+ Tregs within an inflammatory infiltrate (original magnification ×40). F, Colocalization of DCs and CCL22 in inflamed liver tissue. Serial liver sections from a patient with PBC were stained with CD11c (top panels) and CCL22 (bottom panels). Cells that coexpress CD11c and CCL22 can be determined by comparing staining of CD11c (top right panel) and CCL22 (bottom right panel) at higher power (original magnification ×200). Staining with isotype-matched control Abs was negative (left panels, original magnification ×100.) G, Colocalization of DCs and CCL22 in inflamed liver tissue. Staining of FITC-labeled CCL22 (green) with Texas Red-labeled CD11c (red). Areas of coexpression (top left panel) appear yellow. Nuclei were stained with DAPI (blue). Control Abs demonstrated minimal background tissue fluorescence. The findings are representative of the pattern of immunofluorescence staining seen in tissue sections from four donors each for PBC, ALD, AIH, and HCV (original magnification × 200). H, De novo synthesis of CCR4 ligands CCL17 and CCL22 by liver-derived DCs was confirmed by RT-PCR. LIDCs isolated from liver tissue from patients with AIH and PBC were analyzed unstimulated or after stimulation with LPS. Monocyte-derived TNF-α–activated DCs (MoDCs) are shown as a positive control, and unstimulated and stimulated HSECs were used as a negative control. CCL17 and CCL22 mRNA was detected in LIDCs. Representative data from one of three experiments is shown. I, Western blot analysis of CCL17 in liver tissue lysates from different chronic liver diseases and unstimulated or TNF-α/IFN-γ–stimulated HSEC lysates. CCL17 is present in chronically inflamed diseased livers, but it could not be detected in NL tissue lysates or in lysates of HSECs. One representative blot of five experiments is shown. HSEC Unstim, unstimulated HSECs; HSEC Stim, HSECs stimulated with TNF-α and IFN-γ; NANB, non-A non-B seronegative hepatitis; NASH, nonalcoholic steatohepatitis; PSC, primary sclerosing cholangitis.

Close modal

We were unable to detect CCL17 and CCL22 in NL tissue, but both chemokines were detected in chronically inflamed liver tissue by Western blotting (Fig. 6I) and immunohistochemistry (Fig. 6Cii, 6D, 6F). Immunohistochemistry demonstrated that CCL22 is present in cells with the morphology of DCs in inflamed liver tissue (Fig 6Cii, 6D, 6F). Staining was most marked at areas of active inflammation, including portal infiltrates, interface hepatitis, and lobular hepatitis within the parenchyma (Fig. 6Cii, 6D). There was no detectable staining of DCs with isotype-matched control Abs (Fig. 6Ci). Confocal staining showed that intrahepatic CCL22+ DCs were closely associated with LITregs in PTs and inflamed parenchyma (Fig. 6E, Fig. 7) and in close proximity to CD4 and CD8 T cells (Fig. 7). Serial liver sections (Fig. 6F) and dual staining (Fig. 6G) suggested that CCL22 expression was detected on CD11c DCs in inflamed liver.

FIGURE 7.

Close proximity of Tregs, T effectors, and DCs in parenchyma and PT of normal and inflamed liver. A, Quadruple-color confocal images of inflamed liver PT and parenchyma (P) showing LITregs in chronically inflamed liver (AIH liver; top left panel) as an example compared with NL (top right panel). Lower panels show the magnification of parenchyma (P) and PT areas (stains: FoxP3 Tregs, TRITC [red]; CD8, Cy5 [blue]; CD11c, FITC [green]; cell nuclei, DAPI [gray]) (original magnification: upper panels, ×25; lower panels, ×40). B, Quadruple-color confocal microscopy images of CD11c+ DCs and T effector cells (CD4 and CD8) in inflamed PT and parenchyma (P) in a patient with autoimmune liver disease (left panels) and NL (right panels). CD4 (TRITC [red]) and CD8 (Cy3 [blue]) DCs are detected in close proximity with CD11c+ DCs (FITC [green]) (original magnification: upper panels, ×25; lower panels, ×40). Three-color confocal staining images of autoimmune liver hepatic parenchyma (P) infiltrate (interface hepatitis) (C) and PT (D) showing FoxP3 (TRITC [red]) and CD11c (FITC [green]) cell types are closely associated in infiltrates in the inflamed parenchyma lobules (P) and PT. Cell nuclei are stained with DAPI (gray).

FIGURE 7.

Close proximity of Tregs, T effectors, and DCs in parenchyma and PT of normal and inflamed liver. A, Quadruple-color confocal images of inflamed liver PT and parenchyma (P) showing LITregs in chronically inflamed liver (AIH liver; top left panel) as an example compared with NL (top right panel). Lower panels show the magnification of parenchyma (P) and PT areas (stains: FoxP3 Tregs, TRITC [red]; CD8, Cy5 [blue]; CD11c, FITC [green]; cell nuclei, DAPI [gray]) (original magnification: upper panels, ×25; lower panels, ×40). B, Quadruple-color confocal microscopy images of CD11c+ DCs and T effector cells (CD4 and CD8) in inflamed PT and parenchyma (P) in a patient with autoimmune liver disease (left panels) and NL (right panels). CD4 (TRITC [red]) and CD8 (Cy3 [blue]) DCs are detected in close proximity with CD11c+ DCs (FITC [green]) (original magnification: upper panels, ×25; lower panels, ×40). Three-color confocal staining images of autoimmune liver hepatic parenchyma (P) infiltrate (interface hepatitis) (C) and PT (D) showing FoxP3 (TRITC [red]) and CD11c (FITC [green]) cell types are closely associated in infiltrates in the inflamed parenchyma lobules (P) and PT. Cell nuclei are stained with DAPI (gray).

Close modal

To confirm that hepatic DCs secrete CCR4 ligands, we isolated myeloid DCs from normal and inflamed human liver tissue and analyzed them for the expression of CCL17 and CCL22. mRNA (Fig. 6H) and protein (Fig. 6B) for both chemokines were detected in DCs isolated from diseased liver; neither mRNA (Fig. 6H) nor protein (Fig. 6I) could be detected in HSECs, even after cytokine stimulation consistent with the lack of detectable CCL17 and CCL22 on sinusoids by immunohistochemistry (Fig. 6Cii, 6D). Furthermore, we localized staining of CCL22 to LIDCs in liver tissue (Fig. 6G). These data are consistent with previous reports that myeloid DCs secrete CCL17 and CCL22 when stimulated with LPS (26).

We looked for evidence of STAT5 activation in LITregs by staining liver tissue with an Ab that recognizes phosphoSTAT5 (Fig. 8). PhosphoSTAT5 was detected in <20% of the FoxP3+ cells, suggesting that not all FoxP3 Tregs are actively signaling through IL-2R (Fig. 8).

FIGURE 8.

Detection of phosphoSTAT5 in inflamed human livers. A, Single immunohistochemical staining of formalin-fixed, paraffin-embedded liver sections showed phophoSTAT5 staining using mAbs (detected as a brown pigment) in PT and parenchyma (P) of patients with AIH or ALD. Control sections were stained with isotype-matched IgG Abs, and staining was negative (original magnification ×400). B, The percentage of phosphoSTAT5-positive FoxP3 cells in normal and inflamed livers. Paraffin-embedded serial liver sections were stained with FoxP3 mAbs and phosphoSTAT5 Abs; the numbers of cells with positive staining were counted from the same areas, and the percentage positivity of phosphSTAT5 in relation to FoxP3 was calculated. n = 6 each for diseased livers and NL.

FIGURE 8.

Detection of phosphoSTAT5 in inflamed human livers. A, Single immunohistochemical staining of formalin-fixed, paraffin-embedded liver sections showed phophoSTAT5 staining using mAbs (detected as a brown pigment) in PT and parenchyma (P) of patients with AIH or ALD. Control sections were stained with isotype-matched IgG Abs, and staining was negative (original magnification ×400). B, The percentage of phosphoSTAT5-positive FoxP3 cells in normal and inflamed livers. Paraffin-embedded serial liver sections were stained with FoxP3 mAbs and phosphoSTAT5 Abs; the numbers of cells with positive staining were counted from the same areas, and the percentage positivity of phosphSTAT5 in relation to FoxP3 was calculated. n = 6 each for diseased livers and NL.

Close modal

Tregs operate at sites of inflammation in peripheral organs as well as in secondary lymphoid tissues (12, 27). At inflammatory sites, they mediate bystander and Ag-specific suppression of local immune responses to prevent collateral tissue damage in response to injury and infection (2830). The outcome of chronic inflammation is the result of a balance between pro- and anti-inflammatory responses in which Tregs maintain stable chronic inflammation while preventing fulminant destructive inflammation (31). This is particularly important in liver disease, in which intrahepatic Tregs are detected in chronic hepatitis from a variety of causes (5, 32, 33). The frequency and function of intrahepatic Tregs were reported to affect the outcome of chronic infection with hepatitis B and C viruses and to be associated with suppression of the antitumor immune response in hepatocellular carcinoma (4, 33).

We found low numbers of Tregs in nondiseased human liver and a higher number in the liver of patients with a variety of autoimmune and chronic inflammatory liver diseases. Approximately 5% of the CD3 T cells in NL tissue expressed FoxP3 by immunohistochemistry; this proportion was significantly greater in chronic inflammatory liver diseases. We also isolated CD4+CD25+FoxP3+CD127low cells from explanted liver tissue and again found that NL contained low frequencies of Treg and that the frequencies were increased in chronic inflammatory liver diseases. The proportion of liver-derived T cells that stained CD4+CD25+FoxP3+CD127low after isolation from liver tissue was lower than the proportion seen in situ using immunohistochemistry with Abs to FoxP3. This could be because FoxP3 stains other cells in liver tissue, although this seems unlikely because the majority of FoxP3+ liver-derived lymphocytes were CD4+CD25+CD127low. Alternatively, our tissue-isolation methods may not remove all of the LITregs. However, although the proportions of Tregs varied using the two methods, the overall results were similar, with the highest frequencies detected in patients with fulminant seronegative (non-A non-B) hepatitis, the most florid and destructive form of hepatitis that leads to liver failure. These findings suggest that T cell-mediated hepatitis and liver destruction are not a consequence of the failure of Tregs to infiltrate the liver but rather are due to an inability of the recruited cells to control the effector arm of the response.

Analysis of the distribution of Tregs in liver tissue demonstrated them to be closely associated with CD11c+ DCs and CD8 T cells at sites of inflammation in PTs and in the hepatic lobules. Thus, Tregs are ideally situated to mediate suppression of intrahepatic immune responses. To confirm that they were functional, we isolated CD4+CD25+FoxP3+CD127low cells from liver tissue and showed that they suppressed T cell activation in vitro. These cells were CD27high, consistent with previous reports that CD27 expression characterizes tissue-infiltrating Tregs in inflamed synovial tissue1 (13).

The migration and positioning of leukocytes in tissues during homeostasis and in response to inflammation are controlled by the coordinated expression of adhesion molecules and chemokine receptors that determine where and when cells will be recruited via endothelium, as well as their subsequent migration and positioning in tissue. Little is known about the molecular control of Treg recruitment to inflamed tissues, although this will be critical in defining their role in disease and in developing strategies for their therapeutic use. To determine the receptors responsible for the recruitment of Tregs to human liver, we compared chemokine receptor expression on liver-derived Tregs (analyzed without expansion straight from liver tissue) with blood-derived Tregs isolated from the same patient at the time of liver transplantation. In particular, two receptors were overexpressed on liver-infiltrating cells: CXCR3 and CCR4; the majority of LITregs in chronic inflammatory liver diseases express both receptors.

CXCR3 was reported on circulating human Tregs (34, 35), where its expression is presumed to favor homing to inflamed tissue sites. Our data confirm that this is the case because the majority of Tregs in inflamed human liver expressed high levels of functional CXCR3, which promoted their transmigration across liver endothelium under flow. We found that only 15–20% of Tregs in NL expressed CXCR3, which further supports its particular role in the recruitment of Tregs to sites of inflammation. Levels of the CXCR3 ligands are low to undetectable in NL, but they are markedly increased in inflammatory liver diseases (18, 36).

The CXCR3 ligands CXCL9, CXCL10, and CXCL11 are increased in chronic inflammatory liver disease as a consequence of secretion by hepatocytes, stromal cells, and endothelium. They can be readily detected on the glycocalyx of sinusoidal vessels (3739), where they promote the transendothelial migration of effector T cells (18). We developed an in vitro flow-based adhesion assay in which primary cultures of HSECs are grown in microslides under physiological levels of shear stress to model the hepatic sinusoid (18). Treatment of HSECs with TNF-α and IFN-γ induces the expression of CXCL9, CXCL10, and CXCL11, as well as the adhesion molecules ICAM-1 and VCAM-1, thereby recapitulating the phenotype of chronically inflamed sinusoids in vivo (18). We used this assay in the current study to determine the role of CXCR3 on LITregs. We found that signaling through CXCR3 activated integrin-mediated arrest from flow and promoted the transendothelial migration of Tregs across inflamed hepatic sinusoidal endothelium. The fact that a similar degree of inhibition to both processes was seen when a blocking CXCR3 Ab or the Gα protein inhibitor PTx was used suggests that CXCR3 is the dominant chemokine receptor involved. All three CXCR3 ligands were involved in the in vitro flow assays, because blocking individual chemokines had a lesser effect than blocking the common receptor CXCR3. These findings are consistent with studies in murine models of inflammatory disease that report the requirement for CXCR3 expression to allow Tregs to be recruited to inflamed tissues in graft-versus-host disease (40) and to the brain in experimental autoimmune encephalomyelitis (41). Recently, CXCR3 has been implicated in the recruitment of Tregs into the murine liver in response to inflammation and IFN-γ production from NKT cells (42), and Foxp3+ Tregs could upregulate Th1 specifying transcription factor T-bet, which, in turn, promote the expression of the chemokine receptor CXCR3 on Tregs in response to IFN-γ (43). Thus, IFN-inducible chemokines CXCL9 and CXCL10 expressed on inflamed hepatic sinusoidal endothelium promote the recruitment of CXCR3-expressing Tregs under shear stress as they do the recruitment of effector cells (18). A difference between the behavior of effector cells reported in our previous study and Tregs was the involvement of VCAM-1, but not ICAM-1, in the adhesion of Tregs to HSECs, whereas effector cells use ICAM-1 (18).

CCR4 is the other chemokine receptor that was strongly overexpressed on LITregs; it was coexpressed with CXCR3 in nearly 80% of Tregs from chronically inflamed livers. Liver-derived Tregs migrated to the CCR4 ligands CCL17 and CCL22 in vitro, demonstrating that the receptor is functional. CCR4 expression has been associated with Tregs and shown to control migration and immune suppression in the skin (44, 45) and in immunologically tolerant cardiac allografts (46). Until this report, CCR4 was not believed to play a role in the recruitment of T cells to the liver, and few effector cells in the liver express CCR4 (39). CCR4 is the receptor for two chemokines CCL17 and CCL22, both of which are secreted by DCs on maturation, and they recruit and retain Tregs in contact with DCs in lymph nodes (44, 4751). The interaction between Tregs and DCs inhibits DC maturation and the expression of costimulatory molecules required for effector T cell activation (5255). We were unable to detect either chemokine in normal human liver tissue by Western blotting of whole liver tissue or by immunohistochemistry. However, both chemokines were detected in inflamed liver tissue, and immunohistochemistry showed staining restricted to CD11c+ DCs within inflammatory infiltrates in septal areas and lobules. To confirm that LIDCs secrete CCR4 ligands, we isolated myeloid DCs from liver tissue and showed that they expressed and secreted CCL17 and CCL22.

Confocal microscopy of liver tissue revealed that LITregs were closely associated with DCs and CD8 T cells in chronic hepatitis (Fig. 7), suggesting that they are ideally positioned to maintain intrahepatic immune suppression. We propose that localization to these infiltrates occurs in response to sequential chemokine signals. Local proinflammatory cytokines, including IFN-γ, lead to the secretion of CXCL9 and CXCL10 by sinusoidal and parenchyma cells; these chemokines are presented on the glycocalyx of sinusoidal endothelium where they recruit CXCR3high circulating Tregs into the liver. Subsequent migration within the inflamed liver is guided by CCR4, which allows tissue-infiltrating Tregs to respond to CCL17 and CCL22 secreted by activated intrahepatic DCs, thereby resulting in their accumulation within DC-rich inflammatory infiltrates in liver tissue.

This begs the question of why inflammation persists in the presence of so many Tregs. These cells are functional in vitro, but it is possible that their function is suppressed within the liver microenvironment in vivo. To investigate this possibility, we stained the intrahepatic Tregs in situ for phosphoSTAT5. STAT5 phosphorylation is a consequence of IL-2R activation and, thus, provides an indication of whether the intrahepatic Tregs are activated (56). We quantified the numbers of pSTAT5-positive cells in the liver and found that only 5–15% of the liver-infiltrating FOXP3+ cells also expressed pSTAT5. Thus, it is possible that these cells are suppressed in situ, possibly by programmed death receptor-1–dependent mechanisms, as has been proposed for chronic HCV infection (57), allowing the effector arm to drive persistent inflammation.

We thank the staff at the Liver Transplant Unit, Queen Elizabeth Hospital, for their help with sample collection and the patients for donating blood and tissue. We also thank Nicholas Graham, Rebecca Adams, and Janine Youster for help with immunohistochemistry and Gill Muirhead for help with HSEC isolation and culture.

Disclosures The authors have no financial conflicts of interest.

This work was supported by grants from the Medical Research Council (fellowships to Y.H.O., B.E., and L.S.K.W.), Core (fellowship to S.S.), the Wellcome Trust, National Institute for Health Research, and the National Institutes of Health (5RO1AA014257).

Abbreviations used in this paper:

AIH

autoimmune hepatitis

ALD

alcoholic liver disease

DC

dendritic cell

FoxP3

forkhead family transcriptional regulator box P3

HCV

chronic hepatitis C virus

HSEC

hepatic sinusoidal endothelial cell

LIDC

liver-infiltrating dendritic cell

LIL

liver-infiltrating lymphocyte

LITreg

liver-infiltrating regulatory T cell

MoDC

monocyte-derived dendritic cell

NANB

non-A non-B seronegative hepatitis

NASH

nonalcoholic steatohepatitis

NL

normal liver (donor, control)

P

parenchyma

PBC

primary biliary cirrhosis

PBTreg

peripheral blood regulatory T cell

phosphoSTAT5

phosphorylated STAT5

PSC

primary sclerosing cholangitis

PT

portal tract

PTx

pertussis toxin

Treg

regulatory T cell

TRITC

tetramethylrhodamine isothiocyanate.

1
Sakaguchi
S.
,
Powrie
F.
.
2007
.
Emerging challenges in regulatory T cell function and biology.
Science
317
:
627
629
.
2
Crispe
I. N.
,
Giannandrea
M.
,
Klein
I.
,
John
B.
,
Sampson
B.
,
Wuensch
S.
.
2006
.
Cellular and molecular mechanisms of liver tolerance.
Immunol. Rev.
213
:
101
118
.
3
Rushbrook
S. M.
,
Hoare
M.
,
Alexander
G. J.
.
2007
.
T-regulatory lymphocytes and chronic viral hepatitis.
Expert Opin. Biol. Ther.
7
:
1689
1703
.
4
Unitt
E.
,
Rushbrook
S. M.
,
Marshall
A.
,
Davies
S.
,
Gibbs
P.
,
Morris
L. S.
,
Coleman
N.
,
Alexander
G. J.
.
2005
.
Compromised lymphocytes infiltrate hepatocellular carcinoma: the role of T-regulatory cells.
Hepatology
41
:
722
730
.
5
Lan
R. Y.
,
Cheng
C.
,
Lian
Z. X.
,
Tsuneyama
K.
,
Yang
G. X.
,
Moritoki
Y.
,
Chuang
Y. H.
,
Nakamura
T.
,
Saito
S.
,
Shimoda
S.
, et al
.
2006
.
Liver-targeted and peripheral blood alterations of regulatory T cells in primary biliary cirrhosis.
Hepatology
43
:
729
737
.
6
Longhi
M. S.
,
Hussain
M. J.
,
Mitry
R. R.
,
Arora
S. K.
,
Mieli-Vergani
G.
,
Vergani
D.
,
Ma
Y.
.
2006
.
Functional study of CD4+CD25+ regulatory T cells in health and autoimmune hepatitis.
J. Immunol.
176
:
4484
4491
.
7
Sakaguchi
S.
2004
.
Naturally arising CD4+ regulatory T cells for immunologic self-tolerance and negative control of immune responses.
Annu. Rev. Immunol.
22
:
531
562
.
8
Liu
W.
,
Putnam
A. L.
,
Xu-Yu
Z.
,
Szot
G. L.
,
Lee
M. R.
,
Zhu
S.
,
Gottlieb
P. A.
,
Kapranov
P.
,
Gingeras
T. R.
,
Fazekas de St Groth
B.
, et al
.
2006
.
CD127 expression inversely correlates with FoxP3 and suppressive function of human CD4+ T reg cells.
J. Exp. Med.
203
:
1701
1711
.
9
Seddiki
N.
,
Santner-Nanan
B.
,
Martinson
J.
,
Zaunders
J.
,
Sasson
S.
,
Landay
A.
,
Solomon
M.
,
Selby
W.
,
Alexander
S.I.
,
Nanan
R.
, et al
.
2006
.
Expression of interleukin (IL)-2 and IL-7 receptors discriminates between human regulatory and activated T cells
.
J. Exp. Med.
203
:
1693
1700
.
10
Sakaguchi
S.
2005
.
Naturally arising Foxp3-expressing CD25+CD4+ regulatory T cells in immunological tolerance to self and non-self.
Nat. Immunol.
6
:
345
352
.
11
Sakaguchi
S.
,
Sakaguchi
N.
,
Asano
M.
,
Itoh
M.
,
Toda
M.
.
1995
.
Immunologic self-tolerance maintained by activated T cells expressing IL-2 receptor alpha-chains (CD25). Breakdown of a single mechanism of self-tolerance causes various autoimmune diseases.
J. Immunol.
155
:
1151
1164
.
12
Tang
Q.
,
Bluestone
J. A.
.
2008
.
The Foxp3+ regulatory T cell: a jack of all trades, master of regulation.
Nat. Immunol.
9
:
239
244
.
13
Ruprecht
C. R.
,
Gattorno
M.
,
Ferlito
F.
,
Gregorio
A.
,
Martini
A.
,
Lanzavecchia
A.
,
Sallusto
F.
.
2005
.
Coexpression of CD25 and CD27 identifies FoxP3+ regulatory T cells in inflamed synovia.
J. Exp. Med.
201
:
1793
1803
.
14
Cao
D.
,
Malmström
V.
,
Baecher-Allan
C.
,
Hafler
D.
,
Klareskog
L.
,
Trollmo
C.
.
2003
.
Isolation and functional characterization of regulatory CD25brightCD4+ T cells from the target organ of patients with rheumatoid arthritis.
Eur. J. Immunol.
33
:
215
223
.
15
van Amelsfort
J. M.
,
Jacobs
K. M.
,
Bijlsma
J. W.
,
Lafeber
F. P.
,
Taams
L. S.
.
2004
.
CD4(+)CD25(+) regulatory T cells in rheumatoid arthritis: differences in the presence, phenotype, and function between peripheral blood and synovial fluid.
Arthritis Rheum.
50
:
2775
2785
.
16
Belkaid
Y.
2008
.
Role of Foxp3-positive regulatory T cells during infection.
Eur. J. Immunol.
38
:
918
921
.
17
Strauss
L.
,
Bergmann
C.
,
Szczepanski
M.
,
Gooding
W.
,
Johnson
J. T.
,
Whiteside
T. L.
.
2007
.
A unique subset of CD4+CD25highFoxp3+ T cells secreting interleukin-10 and transforming growth factor-beta1 mediates suppression in the tumor microenvironment.
Clin. Cancer Res.
13
:
4345
4354
.
18
Curbishley
S. M.
,
Eksteen
B.
,
Gladue
R. P.
,
Lalor
P.
,
Adams
D. H.
.
2005
.
CXCR 3 activation promotes lymphocyte transendothelial migration across human hepatic endothelium under fluid flow.
Am. J. Pathol.
167
:
887
899
.
19
Eksteen
B.
,
Miles
A.
,
Curbishley
S. M.
,
Tselepis
C.
,
Grant
A. J.
,
Walker
L. S.
,
Adams
D. H.
.
2006
.
Epithelial inflammation is associated with CCL28 production and the recruitment of regulatory T cells expressing CCR10.
J. Immunol.
177
:
593
603
.
20
Lai
W. K.
,
Curbishley
S. M.
,
Goddard
S.
,
Alabraba
E.
,
Shaw
J.
,
Youster
J.
,
McKeating
J.
,
Adams
D. H.
.
2007
.
Hepatitis C is associated with perturbation of intrahepatic myeloid and plasmacytoid dendritic cell function.
J. Hepatol.
47
:
338
347
.
21
Goddard
S.
,
Youster
J.
,
Morgan
E.
,
Adams
D. H.
.
2004
.
Interleukin-10 secretion differentiates dendritic cells from human liver and skin.
Am. J. Pathol.
164
:
511
519
.
22
Lalor
P. F.
,
Edwards
S.
,
McNab
G.
,
Salmi
M.
,
Jalkanen
S.
,
Adams
D. H.
.
2002
.
Vascular adhesion protein-1 mediates adhesion and transmigration of lymphocytes on human hepatic endothelial cells.
J. Immunol.
169
:
983
992
.
23
Grant
A. J.
,
Goddard
S.
,
Ahmed-Choudhury
J.
,
Reynolds
G.
,
Jackson
D. G.
,
Briskin
M.
,
Wu
L.
,
Hübscher
S. G.
,
Adams
D. H.
.
2002
.
Hepatic expression of secondary lymphoid chemokine (CCL21) promotes the development of portal-associated lymphoid tissue in chronic inflammatory liver disease.
Am. J. Pathol.
160
:
1445
1455
.
24
Lai
W. K.
,
Sun
P. J.
,
Zhang
J.
,
Jennings
A.
,
Lalor
P. F.
,
Hubscher
S.
,
McKeating
J. A.
,
Adams
D. H.
.
2006
.
Expression of DC-SIGN and DC-SIGNR on human sinusoidal endothelium: a role for capturing hepatitis C virus particles.
Am. J. Pathol.
169
:
200
208
.
25
Gramberg
T.
,
Soilleux
E.
,
Fisch
T.
,
Lalor
P. F.
,
Hofmann
H.
,
Wheeldon
S.
,
Cotterill
A.
,
Wegele
A.
,
Winkler
T.
,
Adams
D. H.
,
Pohlmann
S.
.
2007
.
Interactions of LSECtin and DC-SIGN/DC-SIGNR with viral ligands: Differential pH dependence, internalization and virion binding
.
Virology
373
:
189
201
.
26
Penna
G.
,
Vulcano
M.
,
Roncari
A.
,
Facchetti
F.
,
Sozzani
S.
,
Adorini
L.
.
2002
.
Cutting edge: differential chemokine production by myeloid and plasmacytoid dendritic cells.
J. Immunol.
169
:
6673
6676
.
27
Walker
L. S.
,
Chodos
A.
,
Eggena
M.
,
Dooms
H.
,
Abbas
A. K.
.
2003
.
Antigen-dependent proliferation of CD4+ CD25+ regulatory T cells in vivo.
J. Exp. Med.
198
:
249
258
.
28
Ni Choileain
N.
,
MacConmara
M.
,
Zang
Y.
,
Murphy
T. J.
,
Mannick
J. A.
,
Lederer
J. A.
.
2006
.
Enhanced regulatory T cell activity is an element of the host response to injury.
J. Immunol.
176
:
225
236
.
29
Holmén
N.
,
Lundgren
A.
,
Lundin
S.
,
Bergin
A. M.
,
Rudin
A.
,
Sjövall
H.
,
Ohman
L.
.
2006
.
Functional CD4+CD25high regulatory T cells are enriched in the colonic mucosa of patients with active ulcerative colitis and increase with disease activity.
Inflamm. Bowel Dis.
12
:
447
456
.
30
O’Connor
R. A.
,
Malpass
K. H.
,
Anderton
S. M.
.
2007
.
The inflamed central nervous system drives the activation and rapid proliferation of Foxp3+ regulatory T cells.
J. Immunol.
179
:
958
966
.
31
Westendorf
A. M.
,
Templin
M.
,
Geffers
R.
,
Deppenmeier
S.
,
Gruber
A. D.
,
Probst-Kepper
M.
,
Hansen
W.
,
Liblau
R. S.
,
Gunzer
F.
,
Bruder
D.
,
Buer
J.
.
2005
.
CD4+ T cell mediated intestinal immunity: chronic inflammation versus immune regulation.
Gut
54
:
60
69
.
32
Longhi
M. S.
,
Ma
Y.
,
Mitry
R. R.
,
Bogdanos
D. P.
,
Heneghan
M.
,
Cheeseman
P.
,
Mieli-Vergani
G.
,
Vergani
D.
.
2005
.
Effect of CD4+ CD25+ regulatory T-cells on CD8 T-cell function in patients with autoimmune hepatitis.
J. Autoimmun.
25
:
63
71
.
33
Rushbrook
S. M.
,
Ward
S. M.
,
Unitt
E.
,
Vowler
S. L.
,
Lucas
M.
,
Klenerman
P.
,
Alexander
G. J.
.
2005
.
Regulatory T cells suppress in vitro proliferation of virus-specific CD8+ T cells during persistent hepatitis C virus infection.
J. Virol.
79
:
7852
7859
.
34
Lim
H. W.
,
Lee
J.
,
Hillsamer
P.
,
Kim
C. H.
.
2008
.
Human Th17 cells share major trafficking receptors with both polarized effector T cells and FOXP3+ regulatory T cells.
J. Immunol.
180
:
122
129
.
35
Lim
H. W.
,
Broxmeyer
H. E.
,
Kim
C. H.
.
2006
.
Regulation of trafficking receptor expression in human forkhead box P3+ regulatory T cells.
J. Immunol.
177
:
840
851
.
36
Harvey
C. E.
,
Post
J. J.
,
Palladinetti
P.
,
Freeman
A. J.
,
Ffrench
R. A.
,
Kumar
R. K.
,
Marinos
G.
,
Lloyd
A. R.
.
2003
.
Expression of the chemokine IP-10 (CXCL10) by hepatocytes in chronic hepatitis C virus infection correlates with histological severity and lobular inflammation.
J. Leukoc. Biol.
74
:
360
369
.
37
Narumi
S.
,
Tominaga
Y.
,
Tamaru
M.
,
Shimai
S.
,
Okumura
H.
,
Nishioji
K.
,
Itoh
Y.
,
Okanoue
T.
.
1997
.
Expression of IFN-inducible protein-10 in chronic hepatitis.
J. Immunol.
158
:
5536
5544
.
38
Tamaru
M.
,
Nishioji
K.
,
Kobayashi
Y.
,
Watanabe
Y.
,
Itoh
Y.
,
Okanoue
T.
,
Murai
M.
,
Matsushima
K.
,
Narumi
S.
.
2000
.
Liver-infiltrating T lymphocytes are attracted selectively by IFN-inducible protein-10.
Cytokine
12
:
299
308
.
39
Shields
P. L.
,
Morland
C. M.
,
Salmon
M.
,
Qin
S.
,
Hubscher
S. G.
,
Adams
D. H.
.
1999
.
Chemokine and chemokine receptor interactions provide a mechanism for selective T cell recruitment to specific liver compartments within hepatitis C-infected liver.
J. Immunol.
163
:
6236
6243
.
40
Hasegawa
H.
,
Inoue
A.
,
Kohno
M.
,
Lei
J.
,
Miyazaki
T.
,
Yoshie
O.
,
Nose
M.
,
Yasukawa
M.
.
2008
.
Therapeutic effect of CXCR3-expressing regulatory T cells on liver, lung and intestinal damages in a murine acute GVHD model.
Gene Ther.
15
:
171
182
.
41
Müller
M.
,
Carter
S. L.
,
Hofer
M. J.
,
Manders
P.
,
Getts
D. R.
,
Getts
M. T.
,
Dreykluft
A.
,
Lu
B.
,
Gerard
C.
,
King
N. J.
,
Campbell
I. L.
.
2007
.
CXCR3 signaling reduces the severity of experimental autoimmune encephalomyelitis by controlling the parenchymal distribution of effector and regulatory T cells in the central nervous system.
J. Immunol.
179
:
2774
2786
.
42
Santodomingo-Garzon
T.
,
Han
J.
,
Le
T.
,
Yang
Y.
,
Swain
M. G.
.
2009
.
Natural killer T cells regulate the homing of chemokine CXC receptor 3-positive regulatory T cells to the liver in mice
.
Hepatology
49
:
1267
1276
.
43
Koch
M. A.
,
Tucker-Heard
G.
,
Perdue
N. R.
,
Killebrew
J. R.
,
Urdahl
K. B.
,
Campbell
D. J.
.
2009
.
The transcription factor T-bet controls regulatory T cell homeostasis and function during type 1 inflammation.
Nat. Immunol.
10
:
595
602
.
44
Sather
B. D.
,
Treuting
P.
,
Perdue
N.
,
Miazgowicz
M.
,
Fontenot
J. D.
,
Rudensky
A. Y.
,
Campbell
D. J.
.
2007
.
Altering the distribution of Foxp3(+) regulatory T cells results in tissue-specific inflammatory disease.
J. Exp. Med.
204
:
1335
1347
.
45
Iellem
A.
,
Mariani
M.
,
Lang
R.
,
Recalde
H.
,
Panina-Bordignon
P.
,
Sinigaglia
F.
,
D’Ambrosio
D.
.
2001
.
Unique chemotactic response profile and specific expression of chemokine receptors CCR4 and CCR8 by CD4(+)CD25(+) regulatory T cells.
J. Exp. Med.
194
:
847
853
.
46
Lee
I.
,
Wang
L.
,
Wells
A. D.
,
Dorf
M. E.
,
Ozkaynak
E.
,
Hancock
W. W.
.
2005
.
Recruitment of Foxp3+ T regulatory cells mediating allograft tolerance depends on the CCR4 chemokine receptor.
J. Exp. Med.
201
:
1037
1044
.
47
Katou
F.
,
Ohtani
H.
,
Nakayama
T.
,
Ono
K.
,
Matsushima
K.
,
Saaristo
A.
,
Nagura
H.
,
Yoshie
O.
,
Motegi
K.
.
2001
.
Macrophage-derived chemokine (MDC/CCL22) and CCR4 are involved in the formation of T lymphocyte-dendritic cell clusters in human inflamed skin and secondary lymphoid tissue.
Am. J. Pathol.
158
:
1263
1270
.
48
Curiel
T. J.
,
Coukos
G.
,
Zou
L.
,
Alvarez
X.
,
Cheng
P.
,
Mottram
P.
,
Evdemon-Hogan
M.
,
Conejo-Garcia
J. R.
,
Zhang
L.
,
Burow
M.
, et al
.
2004
.
Specific recruitment of regulatory T cells in ovarian carcinoma fosters immune privilege and predicts reduced survival.
Nat. Med.
10
:
942
949
.
49
Cardoso
C. R.
,
Garlet
G. P.
,
Moreira
A. P.
,
Júnior
W. M.
,
Rossi
M. A.
,
Silva
J. S.
.
2008
.
Characterization of CD4+CD25+ natural regulatory T cells in the inflammatory infiltrate of human chronic periodontitis.
J. Leukoc. Biol.
84
:
311
318
.
50
Tang
H. L.
,
Cyster
J. G.
.
1999
.
Chemokine up-regulation and activated T cell attraction by maturing dendritic cells.
Science
284
:
819
822
.
51
Vulcano
M.
,
Albanesi
C.
,
Stoppacciaro
A.
,
Bagnati
R.
,
D’Amico
G.
,
Struyf
S.
,
Transidico
P.
,
Bonecchi
R.
,
Del Prete
A.
,
Allavena
P.
, et al
.
2001
.
Dendritic cells as a major source of macrophage-derived chemokine/CCL22 in vitro and in vivo.
Eur. J. Immunol.
31
:
812
822
.
52
Sansom
D. M.
,
Walker
L. S.
.
2006
.
The role of CD28 and cytotoxic T-lymphocyte antigen-4 (CTLA-4) in regulatory T-cell biology.
Immunol. Rev.
212
:
131
148
.
53
Wing
K.
,
Onishi
Y.
,
Prieto-Martin
P.
,
Yamaguchi
T.
,
Miyara
M.
,
Fehervari
Z.
,
Nomura
T.
,
Sakaguchi
S.
.
2008
.
CTLA-4 control over Foxp3+ regulatory T cell function.
Science
322
:
271
275
.
54
Houot
R.
,
Perrot
I.
,
Garcia
E.
,
Durand
I.
,
Lebecque
S.
.
2006
.
Human CD4+CD25high regulatory T cells modulate myeloid but not plasmacytoid dendritic cells activation.
J. Immunol.
176
:
5293
5298
.
55
Bayry
J.
,
Triebel
F.
,
Kaveri
S. V.
,
Tough
D. F.
.
2007
.
Human dendritic cells acquire a semimature phenotype and lymph node homing potential through interaction with CD4+CD25+ regulatory T cells.
J. Immunol.
178
:
4184
4193
.
56
O’Gorman
W. E.
,
Dooms
H.
,
Thorne
S. H.
,
Kuswanto
W. F.
,
Simonds
E. F.
,
Krutzik
P. O.
,
Nolan
G. P.
,
Abbas
A. K.
.
2009
.
The initial phase of an immune response functions to activate regulatory T cells.
J. Immunol.
183
:
332
339
.
57
Franceschini
D.
,
Paroli
M.
,
Francavilla
V.
,
Videtta
M.
,
Morrone
S.
,
Labbadia
G.
,
Cerino
A.
,
Mondelli
M. U.
,
Barnaba
V.
.
2009
.
PD-L1 negatively regulates CD4+CD25+Foxp3+ Tregs by limiting STAT-5 phosphorylation in patients chronically infected with HCV.
J. Clin. Invest.
119
:
551
564
.