The majority of chronic active hepatitis is immune-mediated liver injury (9). Many investigators have reported Treg frequency variation in the peripheral blood in acute liver injury, chronic liver diseases, and liver cancer, but there are limited data on intrahepatic Treg. Reduction in CD4⁺CD25^highCD127^low Treg frequency has been described in patients with alcoholic hepatitis (10). Progression from non-alcoholic fatty liver to non-alcoholic steatohepatitis is characterized by a higher frequency of Th17 cells in the liver and an increased ratio of Th17/resting CD4⁺CD45RA⁺CD25^high Treg in peripheral blood (11). We, and others, have reported that there is an increase in Treg frequency in parallel with effector immune cells in autoimmune liver diseases (AILD) (12–15). Treg also appear to play a role in the immunopathogenesis of primary biliary cholangitis (PBC) (16). Indeed, reduced FoxP3 expression in Treg has been described in the portal tracts of patients with PBC (17). Our group has previously reported the existence of a gut–liver link with the aberrant homing of mucosal T cells from the gut to the liver and extra-intestinal manifestations being seen in inflammatory bowel disease (18–20). Biliary epithelial inflammation has also been associated with the accumulation of CCR10-expressing Treg around the bile ducts in the liver (21).

In the setting of acute liver injury, such as acute viral hepatitis A, the size of the Treg pool was contracted due to Treg apoptosis via a Fas-mediated mechanism (22). Hepatitis B (HBV) pathogenesis is immunologically mediated and increased frequencies of CD4⁺ CD25^highCD45RO⁺ Treg and cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4) cells were noted in the peripheral blood of patients compared with controls and in patients who had recovered from a previous episode of HBV infection (23, 24). However, in HBV-related acute or chronic liver failure, while there was a reduction noted in CD4⁺ T cells, Treg numbers remained unchanged (25). In addition, serial biopsies from patients chronically infected with hepatitis C virus, taken during and after antiviral therapy, suggested that intrahepatic CD4⁺CD25^highFOXP3⁺ Treg frequencies were increased upon interferon and ribavirin therapy in about half of patients, indicating stronger regulation of intrahepatic immunity by Treg during antiviral therapy (26). It is generally accepted that Treg are not beneficial in the setting of liver cancer as an increased Treg frequency correlates with CD8⁺ T cell impairment and poor survival of patients (27). All this body of evidence suggests that Treg play a major role in different types of liver diseases.

The Treg population has been classified as CD25⁺CD45RA⁺FOXP3^low resting, CD25⁺⁺CD45RA^negFOXP3^high activated, suppressive Treg and CD25⁺CD45RA^negFOXP3^low cytokine-secreting non-suppressive Treg (28) (Figure 1). We recently described intrahepatic Treg as predominantly effector memory lymphocytes with PD1^low, CD69^high phenotype (15) and expressing both hepatic homing CXCR3 and biliary tropic CCR6 chemokine receptors (14) (Figure 1).

Regulatory T cells are potent mediators of self-tolerance in the periphery and function via multiple mechanisms to achieve immune modulation. Treg exert their functions by (i) inhibiting the function or maturation of antigen-presenting cells (APCs), (ii) destroying target cells by inducing apoptosis, (iii) causing metabolic disruption via the adenosine pathway, and (iv) by secreting immunosuppressive cytokines, transforming growth factor beta (TGF-β), and IL-10 or competitive consumption of survival cytokines in particular IL-2 (Figure 2).

T-lymphocyte-associated antigen 4 (CTLA-4) protein expression on Treg plays an important role in the suppressor function of Treg (29). In particular, deficiency of CTLA-4 in Treg impairs their suppressive function to result in fatal T cell-mediated autoimmune disease. CTLA-4 via the process of trans-endocytosis depletes its two ligands, CD80 and CD86 from dendritic cells (DCs), thereby removing their availability to act as costimulatory ligands through CD28 (30) (Figure 2). In vitro studies have also identified that CTLA-4 can additionally suppress T-cell proliferation via upregulation of several essential amino acid consuming enzymes, including indolamine 2,3-dioxygenase, histidine ammonia lyase, nitric oxide synthase 2, and l-threonine 3-dehydrogenase in APC through binding CD80 on the APC cell surface (31). IDO-positive DCs represent a regulatory subset of APCs in humans (32) (Figure 2). Regulation of tryptophan metabolism by indolamine 2,3-dioxygenase (IDO) in DCs is a highly versatile modulator of immunity. IDO converts tryptophan to kynurenine, which inhibits the proliferation of effector T cells (33) (Figure 2) but promotes FoxP3 induction in a mechanism involving reduced PI3K/mTOR signaling (31). Importantly, we have shown that human intrahepatic Treg reside close to DCs and effector T cells to exert their suppressive function in the areas of chronic hepatitis, either in the lobules or in the areas of interface hepatitis (34).

TGF-β1, an immunosuppressive cytokine, supports the maintenance of FoxP3 expression, regulatory function, and homeostasis in peripheral Treg (35). It is also a crucial cytokine, which dictates the development of a Treg versus Th17 lineage (36) and limits T effector proliferation via induction of essential amino acid catabolizing enzymes in APCs. IL-10 released by Treg also promotes essential amino acid depletion via upregulation of certain catabolizing enzymes in APCs (31). The liver is enriched with TGF-β1 (37, 38) (Figure 3). Recently, CD103⁺ intestinal DCs have been shown to promote a tolerogenic environment via integrin αvβ8-mediated activation of TGF-β (39), and DCs lacking αvβ8 fail to induce Treg (40).

Treg could also suppress or kill responder T cells after cell to cell contact in a perforin-dependent and independent manner by granzymes A and B (41, 42). Treg can also inhibit T cell proliferation by inducing pericellular adenosine from extracellular nucleotides (43, 44), which is catalyzed by ectoenzymes, CD39, and CD73 expressed on Treg (45, 46). The coordinated expression of CD39/CD73 on Treg and the adenosine A2A receptor on activated T effector cells generates immunosuppressive loops, implementing the inhibitory function of Treg cells (43, 47) (Figure 2). We reported previously that there is an increase in the frequency of intrahepatic Treg in inflamed human livers (34) in parallel with total CD4 T cells, and they had high expression of their functional markers such as CTLA-4, CD39, and also secreted IL-10.

IL-2 is a potent inducer of T-cell proliferation and T-helper cell differentiation and is necessary for the survival and function of memory T cells. It is also important for the development, survival, and function of Treg (45, 48, 49). Treg have higher expression of the high-affinity IL-2 receptor subunit, CD25, than effector T cells; thus, Treg can suppress effector cell proliferation and differentiation by competitive consumption of IL-2 when located in close proximity (50). However, the functional capacity of intrahepatic Treg is reduced in the IL-2-deficient inflamed hepatic microenvironment (15). We demonstrated that the main source of intrahepatic IL-2 is activated T effector cells, and this cytokine is crucial for Stat5 signaling in Treg (15). IL-2 is also required for the homeostatic maintenance of Treg. Taken together, such low intrahepatic IL-2, which appears reduced further in disease, might contribute to a lack of Treg-conferred protection and the progression of immune-mediated liver diseases. Indeed, autoimmune diseases can be induced by IL-2 neutralization or defective IL-2 production (51, 52).

Thus, Treg exert suppression in multiple pathways. The mechanism is dependent on the specific tissue and its microenvironment (Figure 2).

Intrahepatic Treg are residing in a microenvironment, which is deprived of oxygen and enriched with microbes, metabolic products; inflammatory cytokines, and hormones, which all play a role in Treg differentiation and function (Figure 3). The liver is enriched with fat-soluble vitamins. All-Trans retinoic acid (ATRA), a metabolite of vitamin A, is enriched in Ito cells (stellate cells) and plays a pivotal role in maintaining the stability and function of Treg in the inflammatory milieu (53, 54). Retinoic acid can also directly promote TGF-β-mediated conversion of naive T cells to cells of FoxP3⁺ Treg phenotype (55, 56). In addition, retinoic acid can prime human DCs to induce gut homing, IL-10-producing Treg (57) (Figure 3).

The active form of vitamin D, 1,25-dihydroxyvitamin D3 and IL-2 synergistically combine to inhibit T cell production of inflammatory cytokines and promote development of Treg expressing CTLA-4 and FoxP3 (58) (Figure 3). One recent report showed that short-chain fatty acids (SCFA) from gut microbiota-derived bacterial fermentation products regulate the size and function of the colonic Treg pool and protect against colitis, suggesting that abundant microbial metabolites underlie adaptive immune microbiota coadaptation and promote colonic homeostasis (59). We have also reported that the inflamed liver microenvironment is enriched with pro-inflammatory cytokines IL-1, IL-12, IL-6, IL-8, and TNFα, but deprived of the Treg survival cytokine IL-2 (15). Nonetheless, although they present a somewhat reduced regulatory capacity, intrahepatic Treg still possess an intact functional capacity and maintain their lineage for a short time in culture conditions that mimic the intrahepatic environment (15).

There is a shift in metabolic supply-and-demand ratios during inflammation. Tissue hypoxia within inflammatory lesions dictates an anti-inflammatory program by driving expression of hypoxia-inducible factor (HIF)-1α that acts to increase the frequency and suppressive properties of thymic Treg (60, 61). Indeed, hypoxic Treg were more effective than normoxic cells in suppressing the proliferation of effector lymphocytes (61). The hepatic microenvironment is a hypoxic atmosphere especially around zone 3, where cells are closer to the central vein and most distant from the hepatic artery oxygen supply. Thus, while not proven, a gradient in Treg potency might be expected through the liver.

The liver is continuously exposed to gut microbes and bacterial toxins via its portal venous flow. It is clear that the microbiome has a strong influence on the immune system. For example, breast-fed infants develop robust populations of memory T cells as well as T helper 17 (Th17) cells within the memory pool, whereas bottle-fed infants do not, and this may partly explain the variation in human susceptibility to conditions with an immune basis, as well as the variable protection against certain infectious diseases (62). Oral bacteria administration in mice also promotes Treg and alleviates bowel inflammation in a model of immune-mediated colitis (63); thus, the microbiome may serve as a target for future Treg-based immunotherapies.

Leukocyte trafficking and positioning within tissues is directed by chemokines. Thus, chemokines play critical roles in regulating immune responses and inflammation (64, 65). Chemokines can be classified into “inflammatory” and “homeostatic/constitutive” based on whether they are induced by inflammation or constitutively expressed and involved in homeostatic immune regulation (64). Hepatic Treg express a unique range of chemokine receptors, which interact with corresponding chemokines, and these receptors are crucial for their homing and positioning in the inflammatory liver tissues. Human intrahepatic Treg express the chemokine receptor CXCR3 for recruitment across hepatic sinusoids, CCR4 for positioning close to hepatic DCs and both CXCR3 and CCR6 chemokine receptors for positioning around bile ducts (15, 34). This chemokine receptor expression profile is essential for Treg to locate at the site of inflammation and to interact with other immune cells. CXCR3 deficiency has been shown to exacerbate liver disease and abrogates tolerance in mouse models of immune-mediated hepatitis (66).

As has been discussed, Treg play a pivotal role in controlling the magnitude of immune responses to provide tolerance to self-antigens and to limit tissue damage caused by immune activation in response to innocuous antigens. Given the contribution of aberrant immune control in the progression of disease including: (1) suppressed effector immune response which results in unwanted Treg activity in cancer, (2) impaired immune-regulatory function in autoimmune or inflammatory diseases, and (3) the deleterious consequences of life-long immunosuppression therapy following organ transplantation, the prospect to control disease progression through targeting the regulatory cells in settings of cancer, solid organ and hematopoietic cell transplantation, transplant rejection, and autoimmune diseases has been an attractive option for clinicians over many years.

The cytokine IL-2 is essential for the function and expansion of Treg (49). Thus, to improve immune regulation through low-dose, IL-2 therapy, which targets the Treg selectively in contrast to effector, has been tested with positive outcomes in human autoimmune-related diseases in Phase I and II settings (67–70).

The establishment over recent years of good manufacturing practice (GMP)-compliant reagents and equipment that can allow the isolation of cells according to their expression profile of certain surface proteins has made it possible to isolate Treg from the peripheral circulation as a cell immunotherapy. The concept underlying this “Regulatory T cell therapy” is that administering a concentrated source of a desired population of Treg can tip the balance of the patient’s immune system to enhance its regulatory capacity. It builds on the knowledge that a lack of Treg function due to mutation of the Treg lineage-defining transcription factor FOXP3 leads to the X-linked autoimmune syndrome immune dysregulation, polyendocrinopathy, enteropathy X-linked syndrome (IPEX) (71), and, upon the seminal observation by Sakaguchi and colleagues, that giving CD4⁺CD25⁺ Treg could prevent autoimmune disease (2). In this section, we discuss developments in the field of Treg immunotherapy and its potential to be used to treat liver diseases in the future.

To be used therapeutically, a cellular product must be deemed GMP-compliant. This has massive implications upon its production compared with a basic laboratory reagent as the final product and all the reagents and equipment used in its manufacture must exceed a standard clinical quality (purity) and sterility. Accordingly, all staff involved in its preparation must be highly skilled in aseptic procedures and must receive regular re-training. There must be a detailed log of the production and characterization of the final product. Altogether, this imposes extremely high overheads to the production of cellular therapies and with regard to Treg cell therapy has further limited the precise phenotypic character that can be achieved both for practical and for financial reasons. Consequently, the Treg used in trials until now have differed.

To date, we can define three categories of GMP-grade clinical Treg: first generation (CD4⁺CD25⁺); second generation, bone fide Treg (CD4⁺CD25⁺CD127^low/−), and third generation naive Treg (CD4⁺CD25⁺CD127^low/−CD45Ra⁺) (Figure 4). There are two options to achieve GMP clinical grade Treg, bead-based GMP technology and flow sorting technology. While first generation Treg can be isolated by magnetic bead-based approaches alone, such as CD8 and CD19 depletion followed by CD25 enrichment, there are no GMP bead reagents to deplete CD127-expressing effectors. Isolation of second and third generation Treg that are exclusively CD127^low/− depends on the availability of a GMP-compatible flow sorting facility. Tyto technology (Miltenyi Biotec) now exists that can facilitate the isolation and expansion of a highly pure second (CD4⁺CD25⁺CD127^low/−) and third generation (CD4⁺CD25⁺CD127^low/−CD45Ra⁺) GMP Treg preparation straight from a leukapheresis product within a closed system, reducing the risk of loss of sterility and optimizing the purity of the preparation from the start. Moreover, by including Rapamycin and Retinoic acid in the culture cocktail during subsequent expansion, it is possible to ensure that loss of purity through the outgrowth of any contaminating effector T cells is not of concern. Rapamycin precludes the expansion of effector but not regulatory T cells via differential downstream signaling from the IL-2R (72). Furthermore, additional supplementation with IL-2 ± retinoic acid can help to support Treg proliferation and potentiate the functional phenotype (73), and it is possible to measure the demethylation status of the Treg-specific demethylated region (TSDR) prior to administration of the cell therapy product to the patient to verify purity. Although percentage demethylation boundaries are not yet rigorously defined, such could be implemented during trials to help restrict any adverse events based on known lack of purity.

To date, GMP Treg therapy has not been reported in the treatment of the liver diseases either in preventing rejection of donor organs, preventing GVHD posttransplant, enabling withdrawal of immunosuppressants or overcoming any need for them posttransplant, or in inducing or maintaining states of remission in the pre-transplant setting, but evidence from animal studies (2, 3, 74, 75) and trials of Treg immunotherapy in other human disease backgrounds provide sound rationale for the attempted use of this therapy in liver medicine in the future. This is particularly important in an era when the demand for liver transplants exceeds the availability of matched donor livers. Currently, the first clinical trials are underway, investigating the safety and feasibility of Treg therapy in renal transplantation (76, 77), and it is likely that outcomes of this trial could be applicable to similar solid organs such as the liver.

Taken together, studies using GMP Treg, whether first or second generation, appear promising although clearly the optimization of factors including dose, timing relative to transplant/disease stage, as well as the choice of immunosuppression regimen and the frequency of Treg reapplication are important, and it is likely that each of these will be dictated by the individual situation (prophylactic pre or posttransplant versus autoimmunity/GVHD/rejection), the grade of the individual’s disease, and whether the patient’s disease is in remission (requiring withdrawal of standard medication) or in relapse (requiring settling of active immune reactions as well as withdrawal of standard medications). Given that Cyclosporine reduces Treg number, immunosuppression therapies used together with Treg therapy should be tailored to a rapamycin-based regimen along with Tacrolimus, which increases Treg survival. Patients with active viral infection and previous history of malignancy should be excluded from any Treg clinical trial as Treg would have a negative impact on the disease process.

It seems likely that isolated cells will have to be expanded in order to be of adequate number to be used in therapy. Hence, one caveat will, therefore, be the reliable expansion of autologous Treg cells for all patients since most studies, to date, have reported a few subjects who were enrolled but failed to receive treatment or treatment at the intended dose due to an inadequate cell yield.

The question over the preparation of the Treg also remains. Until now the Treg tested in clinical trials have been polyclonal and, thus, have ability to function through indirect pathways but also specifically due to a number within the mix that have donor antigen specificity (namely alloantigen-specific Treg). With a view to optimize Treg for transplantation therapy, polyclonal and alloantigen-specific Treg have been compared. Alloantigen-specific Treg represent those Treg that are activated by donor-specific antigens and can be expanded from polyclonal Treg populations in mixed lymphocyte reactions with donor APCs. By contrast polyclonal Treg are expanded with anti-CD3/anti-CD28-coated expander beads. Alloantigen-specific T cells can be identified and selected for further expansion based on their induction of activation markers such as CD69 and CD71. The fact that the therapeutic product created by Todo et al. through mixed lymphocyte reaction had greater suppressive ability over recipient T cells activated by donor as opposed to third-party stimulation was clear evidence of underlying allospecific regulatory activity having been raised within the expanded culture, and, as such, using a system to generate greater donor specificity might be advantageous and should allow considerably reduced numbers of Treg to be used (83). The value of alloantigen-specific Treg in transplantation tolerance was cleverly demonstrated in vivo using a humanized mouse model of human skin graft immune rejection. Allogeneic Treg expanded on myeloid DCs from the skin were more effective at limiting dermal and inflammatory phenotypes of skin rejection by infusions of allogeneic T effector cells than polyclonal Treg (74). Various methods have been proposed to generate GMP-compliant donor alloantigen reactive Treg (darTreg) for therapeutic application. Appropriate alternative cell types to incorporate in coculture with recipient CD4⁺CD25⁺ Treg to drive the expansion of darTreg include donor PBMCs, donor monocyte derived DCs, and donor B cells (92) activated on 3T3 fibroblasts expressing CD40L (93). A number of current phase 1 and phase 2 clinical trials propose the application of darTreg as opposed to polyclonal Treg (Table 1). Similarly, in autoimmunity, where models of disease propose the likely hood of only a defined autoantigen, antigen-specific Treg may be the way forward, but this requires knowledge of the offending antigen. In the liver diseases, autoantigens for type II AIH (94) and PBC (95) are known but for type I AIH and PSC are yet to be discovered precluding the benefits of antigen-specific therapy at the present time. Nonetheless isolation of Treg expressing markers of activation such as latency-associated peptide (LAP) and glycoprotein A repetitions predominant (GARP) might help to generate a product with a higher proportion of antigen-specific Treg (Figure 4).

Clinical trials of Treg therapy for solid organ transplantation and autoimmune diseases that are listed on www.ClinicalTrials.gov and currently recruiting are summarized in Table 1.

Overall, Treg hold promise as personalized therapy in the treatment of liver diseases and, in the future, may be applied in the non-transplant as well as post-transplant setting to overcome or minimize the use of broad spectrum immunosuppressant medications that have unfavorable side effects (79). Their mechanisms of function vary at different tissues. While Treg therapy has been applied in other autoimmune disease settings, such as diabetes and SLE, successful application in the treatment of the liver diseases, including autoimmune-related liver patients, may require consideration of the impact of the inflamed liver microenvironment itself which is enriched with cytokines, microbes, and metabolites. Dissecting the clonotype and gene signature would facilitate exploration of antigen specificity, and therapy can be optimized by administering antigen-specific Treg. Notably, it is already clear that the effective proliferation and function of Treg is dependent on IL-2, yet the liver microenvironment is deficient in IL-2 (15). Thus, it is anticipated that Treg therapy for liver diseases would benefit from adjuvant supply of low-dose IL-2, which can selectively potentiate Treg function without establishing global immune activation. It is now an exciting time to conduct Treg cell therapy with manipulation of different cytokines and the microenvironment to achieve a successful and potential cure in patients with AILD.

All authors listed have made substantial, direct, and intellectual contribution to the work and approved it for publication. The views expressed in this paper are those of the author(s) and not necessarily those of the NHS, the NIHR, or the Department of Health.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.