Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
Potential conflict of interest: Nothing to report.
CD4+CD25high regulatory T cells (Tregs) play a critical role in self-tolerance, as seen in murine autoimmunity. Studies on Tregs in human autoimmunity have focused primarily on peripheral blood samples. A study targeting diseased tissue should identify direct relationships between Tregs and autoimmunity. Peripheral blood samples were collected from 91 patients with primary biliary cirrhosis (PBC), 28 immediate relatives, and 41 healthy controls, and Treg frequencies were determined as a percentage of CD4+CD25high T cells in CD4+TCR-αβ+ T cells. A tissue-targeted determination of frequency and distribution of FoxP3+ Tregs was also performed on 90 different liver tissue specimens exhibiting PBC (n = 52), chronic hepatitis C (CHC) (n = 30), and autoimmune hepatitis (AIH) (n = 8). Treg suppression studies were performed on 50 PBC patients and 27 controls. Patients with PBC demonstrated a relative reduction of Tregs compared with controls (P < .0002). Interestingly, a deficiency in CD4+CD25+ Tregs was also found in the daughters and sisters of PBC patients compared with controls (P < .0007). However, functional studies did not reveal a global PBC Treg defect. The level of FoxP3-expressing Tregs was markedly lower in affected PBC portal tracts compared with CHC and AIH (P < .001). In addition, the CD8+T cell/FoxP3+ Treg ratio was significantly higher in livers of late-stage PBC compared with those of CHC (P < .001) and early-stage AIH (P < .001). In conclusion, these data provide support for a genetic modulation of Treg frequency and illustrate the role Tregs play in the loss of tolerance in PBC. (HEPATOLOGY 2006;43:729–737.)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Primary biliary cirrhosis (PBC) is an enigmatic liver disease characterized by the destruction of small intrahepatic bile ducts, portal inflammation, and the presence of antimitochondrial antibodies (AMAs).1, 2 The presence of AMAs and autoreactive T and B cells, in conjunction with the co-occurrence of other autoimmune diseases, characterizes PBC as a typical autoimmune disease.3 Although the etiology of PBC remains obscure, recent data suggest that autoreactive T cell responses play a major role in its pathophysiology.4–7 CD4+CD25high regulatory T cells (Tregs) have been reported to play a key role in the prevention of autoimmune disease in murine models.8–14 These data prompted us to examine the potential role of Tregs in the pathogenesis of PBC.15
A particular subset of CD4+ T cells expressing forkhead box P3 (FoxP3) and constitutively high mean density levels of CD25 had been identified showing regulatory function in human and murine models of autoimmunity. CD4+CD25high Tregs aid in the maintenance of self-tolerance by suppressing the proliferation and cytokine production of autoreactive lymphocytes after specific T cell receptor (TCR) activation by self-antigens in the presence of interleukin 2.15 There is a large body of data that support a requirement for direct cell-to-cell contact in Treg suppression.16, 17 Phenotypically, CD4+CD25high Tregs are anergic in vitro and do not exhibit unique cell surface markers. Most of the known surface markers of these Tregs, including CD25 and glucocorticoid-induced tumor necrosis factor receptor (GITR), are also upregulated upon activation of conventional CD4+CD25high T cells.15 Currently, the best indicator of Treg function is through the intracellular expression of FoxP3, which is also crucial for Treg development.18
Tregs have been found to play a role in a number of human autoimmune diseases.19–23 Most studies, however, have been performed primarily with peripheral blood samples. Direct cell-to-cell contact generally appeared to be a requirement by which Tregs mediate their regulatory function. It seemed reasonable that a more detailed analysis of the frequency of this cell lineage in not only peripheral blood, but also in the target liver tissue of PBC patients, may provide more definitive evidence for a role of Tregs in the pathogenesis of PBC.
Peripheral blood samples were obtained from 91 patients with PBC, 28 relatives of PBC patients, and 41 healthy controls (Table 1). Relatives consisted of otherwise healthy daughters or sisters of PBC patients with no history of any chronic illness. AMAs were detected through the use of recombinant mitochondrial antigens.24 All patients, including AMA-negative subjects, fulfilled the diagnostic criteria of PBC based on internationally accepted standards and were all undergoing ursodeoxycholic acid treatment.2 Patients who did not have fibrosis at the time of liver biopsy are considered as stages I or II PBC according to Ludwig et al.,25 while those showing signs of fibrosis or cirrhosis at histology were considered as stages III or IV (Table 1). Total serum immunoglobulin M (IgM) levels were measured using the Immuno-Tek human IgM EIA Kit (ZeptoMetrix Corporation, Buffalo, NY) (Table 1). FoxP3 and GTR staining were performed on 18 controls with no history of autoimmune disease along with 19 early-stage and 11 late-stage PBC patients. Treg functional studies were carried out on 27 controls and 50 PBC patients. This study was conducted with approval from the institutional review boards at the University of California–Davis School of Medicine and Toyoma and Kyushu Universities and followed the ethical guidelines of the 1975 Declaration of Helsinki and subsequent modifications. All patients gave written informed consent to participate in the study.
Table 1. Subject Data Distribution
Mean Age ± SEM
Stage I-II (%)
Stage III-IV (%)
P < .001.
P < .001 (comparison between relative control vs. control and PBC).
Liver needle biopsies or resected tissues were obtained from 52 patients with PBC registered in our universities and associated hospitals. Each specimen contained more than four portal tracts encompassing interlobular bile ducts (Table 2). All patients who were donors for liver tissues were seropositive for AMAs. As controls, liver needle biopsies or resected tissues from 30 patients with chronic hepatitis C (CHC) and 8 patients with early-stage autoimmune hepatitis (AIH) were also examined. Ninety-one portal tracts were studied from early-stage CHC patients and 38 from those experiencing late-stage CHC. Table 2 summarizes the patient specimen data. The diagnosis of each case was based on reliable clinical and laboratory data and was independently confirmed histologically by two liver pathologists (C. M. C., K. T.). All tissues were fixed in 10% neutral buffered formalin and embedded in paraffin, and more than 10 4-μm-thick serial sections were cut from each paraffin block.
Peripheral blood mononuclear cells (PBMCs) were isolated using Histopaque-1077 (Sigma Chemical Co., St. Louis, MO). After centrifugation, cells were washed with phosphate-buffered saline containing 0.5% bovine serum albumin, and the viability of cells was confirmed using trypan blue dye exclusion. PBMCs were resuspended in staining buffer (0.5% bovine serum albumin, 0.04% EDTA, 0.05% sodium azide in phosphate-buffered saline), divided into 25-μL aliquots, and preincubated with antihuman FcR-blocking reagent (Miltenyi Biotech Inc., Auburn, CA) for 15 minutes at 4°C. They were then washed and stained with FITC-conjugated antihuman TCR-αβ-1 (BD Biosciences, San Diego, CA), phycoerythrin-Cy5–conjugated antihuman CD4 (eBioscience, San Diego, CA), and R-phycoerythrin–conjugated antihuman CD25 (Miltenyi Biotec) for 15 minutes at 4°C. Stained cells were then washed and fixed with 1% paraformaldehyde in phosphate-buffered saline. A FACScan flow cytometer (BD Immunocytometry Systems, San Jose, CA) was used for data acquisition. For FoxP3 expression, intracellular phycoerythrin-conjugated antihuman FoxP3 (clone PCH101) was used, along with FITC-conjugated antihuman CD4 (eBioscience), TRI-COLOR–conjugated CD25 (Caltag Laboratories, Burlingame, CA), and allophycocyanin-conjugated antihuman CD3 (Biolegend, San Diego, CA). GITR expression was examined with FITC-conjugated antihuman GITR (eBioscience), phycoerythrin-conjugated antihuman CD25 (Miltenyi Biotec), allophycocyanin-conjugated antihuman CD3 (Biolegend), and allophycocyanin-Cy7–conjugated antihuman CD4 (Biolegend). Data were analyzed with CELLQUEST (BD Immunocytometry Systems) and FlowJo (Tree Star, Inc., Ashland, OR) softwares. The gain and gates set for analysis of CD25 density were identical for samples from PBC patients and controls.
Immunostaining of Liver Tissues
Single Immunoenzymatic Staining.
To investigate the distribution and frequency of cells expressing FoxP3 and the ratio of CD8+ to CD3+ cells, single immunoenzymatic staining was performed on sections of liver tissues from each patient using the FoxP3 (ab2481; Abcam, Ltd, Cambridge, UK), CD3 (monoclonal antibody; DAKO, Carpinteria, CA) and CD8 (monoclonal antibody; DAKO) antibodies and the microwave procedure (see supplemental material for FoxP3 specificity at the HEPATOLOGY website: www.interscience.wiley.com/jpages/0270–9139/suppmat/index.html).26 In brief, after deparaffinization and standard antigen retrieval via microwave irradiation, all sections were soaked in 3% H2O2for 5 minutes to inhibit endogenous peroxidase, rinsed, and incubated with 5% bovine serum albumin (bovine serum albumin, Sigma Chemical Co.) for 5 minutes to prevent nonspecific staining. Goat polyclonal anti-FoxP3 antibody (Abcam) diluted to 1/50, mouse monoclonal anti-CD3 (DAKO) or anti-CD8 (DAKO) antibodies diluted to 1/200 as mentioned above were respectively applied to the specimens in a plastic moist chamber for 15 minutes under intermittent microwave irradiation (MI-77, Azuyama, Tokyo, Japan; 250 W, 4 s on, 3 s off) followed by incubation at room temperature for an additional 45 minutes. After washing with Tris-buffered saline twice for 5 minutes each time, peroxidase-conjugated polyclonal antibody against goat immunoglobulin (Simple Fine Stain for goat; Nichirei Co, Tokyo, Japan) and peroxidase-conjugated monoclonal antibody against mouse Ig (Envision-PO-plus for mouse; DAKO) were used as secondary antibodies. The incubation conditions for the secondary antibody were the same as those for the primary antibody. After washing with Tris-buffered saline, sections were developed with 3-3′diaminobenzidine (Sigma) and counterstained with hematoxylin.
Double Immunoenzymatic Staining.
To determine the distribution pattern of the FoxP3+ cells that coexpress CD3 in affected portal tracts of PBC as well as CHC, a double immunoenzymatic staining protocol was performed using formalin-fixed, paraffin-embedded representative liver tissue sections from PBC and CHC patients. Briefly, following deparaffinization, standard antigen retrieval and blocking of endogenous peroxidase was performed on each tissue section as described above. FoxP3 antibody was first incubated with sections using the Envision-PO-Plus system (DAKO) and visualized with 3-3′diaminobenzidine (Sigma), and the addition of anti-CD3 was followed by incubation with alkaline-phosphatase conjugated antimouse immunoglobulin reagents (Envision-AP; DAKO). The sections were then visualized with Fast Blue (Vector).
Evaluation of the Frequency of FoxP3-Positive Cells Relative to Other Cell Lineages.
To evaluate and compare the distribution and frequency of cells positive for FoxP3, CD3, and CD8 in PBC, CHC, and AIH, more than 4 small to medium-sized portal tracts from each early-stage case were selected for examination with an optical microscope. The same visual fields as those chosen for early stage cases were examined within the widened fibrous septa of each late stage case. The number of FoxP3-, CD3-, and CD8-positive cells contained within the portal tracts selected in each specimen was counted at a magnification of ×400. To correct for differences in the sizes of the portal tracts, the proportion of FoxP3+Tregs among the total number of CD3+T cells (FoxP3+ T regs/CD3+ T cells × 100)—that is, the CD3 T cell–corrected value for FoxP3—was determined. In addition, to compare the relative prevalence of FoxP3+ Tregs and CD8+ T cells, the CD8+T cells/FoxP3+ Treg ratio was calculated for each portal tract.
Regulatory T Cell Functional Assays
Enriched populations of CD4+CD25+ and CD4+CD25− T cells were isolated using the human magnetic CD4+CD25+ Regulatory T Cell Isolation Kit (Miltenyi Biotec). The purity of isolated CD4+CD25+ Tregs was greater than 95%. The enriched CD4+CD25+ and CD4+CD25− T cells were either cultured alone at 2 × 104 cells/well or cocultured at 4 × 104 cells/well in autologous combinations in triplicate. Cultures were performed in 96-well round-bottom plates previously coated with either media (for control) or antihuman CD3 monoclonal antibody (BD Biosciences) at a concentration of 0.5 μg/mL. The cultures were then incubated at 37°C for 3 days, and each culture was pulsed with 1 μCi thymidine (3H) and harvested after 16 hours of incubation; counts per minute per well was determined by way of scintillation counting. The percent Treg suppression was then calculated.
PBC Patients Demonstrate Decreased Frequencies of Tregs in Peripheral Blood Samples.
We first set out to determine the relative frequencies of Tregs in the peripheral blood of PBC and control subjects as a potential indicator of decreased Treg activity in PBC patients. The frequency of CD4+CD25high Tregs, defined as a percentage of total CD4+/TCR-αβ+ T cells, was ascertained in PBMC samples from each PBC and control subject. A significantly lower mean frequency of CD4+CD25high T cells was observed in PBMCs from the 91 PBC patients compared with that of the 41 control subjects (4.31% vs. 5.67%, respectively; P < .0002) (Fig. 1A-B).
Given that the development and function of Tregs is critically dependent on FoxP3,18 FoxP3+ Treg frequencies as a percent of CD4+ T cells were also determined for 30 PBC patients and 18 controls. Consistent with the CD4+CD25high Treg findings mentioned above, FoxP3 Tregs were also found to be decreased in PBC patients versus control subjects (4.3% vs. 5.9%, respectively; P < .004). However, it should be noted that a distinct positive population of FoxP3 could not be isolated from flow cytometric plots (Fig. 2C), and thus FoxP3+ Tregs were selected based on isotype data. Through the Foxp3 flow cytometic analysis, we also verified that the CD4+CD25high population contained only FoxP3+ cells (Fig. 2A-B), thus proving CD25 to be a reliable marker for Tregs. Because CD25 is a dependable marker of FoxP3+ Tregs, and since it presents a more distinct population compared with the FoxP3+ population, we examined the FoxP3 and GITR expressions of the CD4+CD25high Treg population in the next section (Table 3).
Table 3. FoxP3 MFI of Regulatory T Cells in Patient and Control Groups
NOTE. MFI values were calculated as the mean ± SEM.
Increased FoxP3 and GITR Expression in CD4+CD25+ Regulatory T Cells of Late-Stage PBC Patients Compared With Early-Stage and Control Subjects.
Interestingly, there appears to be a stage-dependent difference in FoxP3 expression within the CD4+CD25high Treg population of PBC patients. Although the density of FoxP3 expression levels in CD4+CD25high Tregs were similar between control and early-stage PBC subjects, the FoxP3 mean fluorescence intensity was higher in the PBMCs of late-stage PBC patients compared with those of early-stage and control subjects (P < .0334 and P < .0026, respectively) (Table 3). The mean density of GITR expression, another marker for Tregs, was also found to be significantly higher in the PBMCs of PBC patients versus controls (P < .0163) (Table 3). However, no stage-related differences were found in GITR expression.
CD4+CD25high Regulatory T Cell Frequencies Do Not Correlate With Disease Stage, Presence of AMA, or IgM Levels.
To determine the role of CD4+CD25high Tregs in the disease progression of PBC, we analyzed the relationship between the frequency of CD4+CD25high Tregs and disease stage. Patients were divided into 2 groups based on disease stage: patients with stage I and II disease (51%) and patients with stage III and IV disease (49%). The mean frequencies of CD4+CD25high T cells were calculated and compared between the two groups. No correlation was found between disease stage and Treg frequency (data not shown).
Serum AMA and elevated IgM levels are both key features of PBC. We therefore attempted to observe any relationships between Treg frequencies and AMA presence or IgM levels. Our data confirm that patient IgM levels were indeed significantly higher than that of normal controls (Table 1). However, no correlation was observed between Treg frequencies and IgM levels (data not shown). Similarly, no significant difference in Treg frequency was observed between AMA-positive and negative patients (data not shown).
PBC Relatives Demonstrate Decreased Frequencies of CD4+CD25high Tregs.
Genetic factors have been implicated in the pathogenesis of human PBC because there is an increased susceptibility of family members for PBC and other autoimmune diseases.27 This prompted us to analyze Treg frequencies in relatives of PBC patients, which consisted of daughters and sisters of patients. There was a significantly lower frequency of CD4+CD25high T cells in PBC relatives compared with control subjects (4.0% vs. 5.3%, respectively. *P < .0005) (Fig. 3A). No significant differences were noted between values in relatives and PBC patients (Fig. 3A). To eliminate possible age-related Treg frequency differences with the younger relatives, PBMCs from 10 age-matched female control subjects were analyzed. Results confirmed that age-matched relatives of PBC patients continue to display a significant decrease in the mean CD4+CD25high Treg frequency compared with normal controls (4.1% vs. 5.7%, respectively; P < .004) (Fig. 3B). There was no correlation between age and Treg frequency (data not shown). When the data for the relatives of PBC patients were further divided into those from daughters and sisters of PBC patients (14 daughters, 9 sisters), no significant difference was found (data not shown).
Regulatory T Cells From PBC Patients and Controls Demonstrate Similar Degrees of Suppressor Function.
Regulatory T cell suppression of effector T cell proliferation in PBC patients (n = 50) and control subjects (n = 27) was compared via coculture assay (Fig. 4). CD4+CD25+ Tregs were cocultured with the same number of autologous CD4+CD25− T cells in the presence of plate-bound anti-CD3 antibodies along with unmixed cultures of CD4+CD25+ and CD4+CD25− cells. Proliferation was measured as counts per minute per well using thymidine incorporation, and the percentage suppression by Tregs of CD4+CD25− T cells was calculated as mentioned above. The results showed no significant differences between suppressive capabilities of Tregs in the PBMCs of PBC and control subjects (data not shown).
Distribution and Frequency of FoxP3+ T Cells in PBC, CHC, and AIH.
In addition to measuring the frequency of Tregs in peripheral blood, FoxP3+ Tregs were also examined in the target liver tissues. As shown in Fig. 5A, FoxP3+ Tregs were scattered among the intensive lymphoid infiltrates localized to the portal tracts in tissues from PBC, CHC, and AIH patients. As expected, there were no detectable FoxP3+ Tregs observed in the liver tissue of normal controls. However, there were markedly fewer FoxP3+ Tregs observed in liver tissues from PBC patients compared with those from patients with CHC and AIH. In addition, coexpression of membranous CD3 and nuclear FoxP3 was confirmed via double immunoenzymatic staining in liver sections of PBC and CHC (Fig. 5B). The frequency of FoxP3+ Tregs relative to CD3+ T cells in the affected portal tracts from PBC patients was significantly lower than similar tissues from patients with CHC and AIH (P < .001) (Table 4). In contrast, in liver tissues from either early- or late-stage disease, CD3+CD8+ T cells were more frequently observed within the infiltrated portal tracts of PBC patients compared with those from CHC or AIH patients. Interestingly, the CD8+ cytotoxic T lymphocytes in PBC patients were found near damaged, intralobular bile ducts, whereas in CHC and AIH, CD8+ T cells were found in the peripheral areas of the portal CD3+ lymphoid cell aggregates (Fig. 6). In addition, the CD8+T cell/FoxP3+ Treg ratio was significantly higher in late-stage disease PBC patients than in patients with CHC (P < .001) and early-stage AIH (P < .001) (Fig. 7A). Furthermore, the CD8+T cell/FoxP3+ Treg ratio was significantly greater in the liver tissue of PBC patients with bile duct damage in portal tracts compared with those without bile duct damage (P < .001) (Fig. 7B).
Table 4. Frequency of FoxP3+ in CD3+ T Cells in PBC, CHC, and AIH
Recent studies have revealed a critical role of CD4+CD25high Tregs in the prevention of autoimmunity and maintenance of self-tolerance. Some studies have demonstrated that the transfer of T cells lacking the CD4+CD25high Treg subset into athymic nude mice results in the development of various T cell–mediated autoimmune diseases.16, 18 In light of these findings, along with other reports of Treg dysfunctions in human autoimmune disease, we investigated the potential role of Tregs in the pathogenesis of PBC.
PBC is immunopathologically characterized by the presence of specific AMAs, elevated levels of IgM, and the presence of autoreactive CD4+ and CD8+ T cells both in the peripheral blood and in the intrahepatic small bile ducts of patients.3, 4, 28 Given that Tregs are well characterized for their ability to suppress the proliferation and activation of autoreactive CD4+ and CD8+ T cells,16 we set out to determine whether a functional defect or deficiency of Tregs may contribute to disease progression in PBC. One factor that must be considered within the context of our findings is the potential effect of ursodeoxycholic acid—a naturally occurring bile acid taken by all PBC patients in this study—on the resulting data.
Our data demonstrate that PBC patients displayed significantly lower frequencies of CD4+CD25high Tregs as percentages of total TCR-αβ+/CD4+ T cells, which may contribute to the breakdown in tolerance in PBC. FoxP3+ Tregs were also found to be decreased in the patient population, although a distinctive FoxP3+ Treg population was difficult to assess. Subsequently, immunohistochemical experiments revealed Foxp3+Tregs to be scattered among CD3+lymphocytes and aggregating around affected intrahepatic bile ducts in PBC. The role of FoxP3+ Tregs in the liver is unclear, although normal human and murine livers did not display a significant population of Tregs. This may mean that tolerance in the liver is not principally maintained by these Tregs, or that they act in a distal manner. On the other hand, the presence of Tregs in diseased livers can be explained by the influx of infiltrating leukocytes, where Tregs may represent a small percentage, and the inflammatory environment. Our findings show that there was a significant paucity of Foxp3+ Tregs as a fraction of CD3+ T cells in liver tissues from early-stage PBC patients compared with those with early-stage CHC, AIH, and late-stage PBC. On the other hand, relative to Foxp3+ Tregs, the number of CD8+T cell infiltrates in patients with early-stage PBC was significantly higher than that found in patients with early-stage CHC, AIH, and late-stage PBC. CD8+ T cells were our primary focus because they are found at the site of actual bile duct damage and hepatocellular necrosis.2 Our studies demonstrated that the portal tracts of PBC tissues expressing bile duct damage or loss demonstrated a significantly greater CD8+T cell/Foxp3+ Treg ratio than those with intact bile ducts. The paucity of Foxp3+ Tregs along with the increased number of CD8+ T cell infiltrates indicates that a deficiency of CD4+CD25+Foxp3+ Tregs may be a key point in the pathogenesis of PBC. The lack of counter-regulation by Tregs could cause exaggerated CD8+ T cell responses resulting in immunopathology with severe consequences such as bile duct damage or loss in PBC.
Functional assays performed in this study demonstrated that despite the decrease in Treg frequency of PBC patients compared with control subjects, they demonstrate comparable global functional capabilities similar to that of Tregs from control subjects. Interestingly, Tregs from late-stage patients exhibit a slightly higher mean density of expression of FoxP3 than early-stage PBC and control subjects, which may be a result of increased activation of these cells due to disease progression.29, 30 The effect of increased cellular FoxP3 on Treg suppression is not clear, but our data suggest that an increase in FoxP3 does not affect suppression. GITR expression was also increased in Tregs from PBC patients compared with those from normal controls, which may also reflect Treg activation. It is plausible that, despite proper global functioning of Tregs, the decreased frequency of Tregs in PBC patients reflects a reduced local Treg repertoire, particularly in the range of autoreactive TCRs. The reduced Treg TCR repertoire may enhance autoimmune responses toward particular autoantigens such as PDC-E2 in PBC.
The frequency of Tregs in the PBMCs of sisters and daughters of PBC patients was also significantly lower than in those of age-matched control subjects (Fig. 2). These results support the view of a genetic component that contributes to the etiology of PBC. This component may precede the occurrence of disease and play an important role if accompanied by the exacerbation of another pathway or insult, such as the involvement of xenobiotics and microorganisms in the generation of autoantibodies through chemical modification or mimicry of host proteins. It should be noted that PBC is one of many autoimmune diseases in which a role for genetics has been previously defined.27 Decreased frequency of Tregs found in PBC patients and their immediate relatives may involve genes controlling the thymic development of this regulatory subset as well as dysregulation in peripheral Treg generation. Several factors have already been implicated in the generation of Tregs, including estrogen and vitamin D3, both of which have been shown to increase Treg frequencies in murine models.31, 32 Studies on the molecular mechanisms involved in regulating the frequency of this subset may shed further light on the role of Tregs in PBC.