Regulatory T cells (Treg) represent a distinct subset of CD4⁺ lymphocytes which play a pivotal role in the maintenance of peripheral tolerance by actively preventing autoimmunity (1). The development and suppressive functions of Treg are essentially regulated by constitutive expression of the transcription factor Forkhead Box 3 (FoxP3) (2). As such, a disruption of the encoding gene gives rise to rampant expansion of autoreactive CD4⁺ T cells, which then infiltrate several organs, exacerbate and perpetuate tissue insult by recruiting other inflammatory cells. The consequence is a lethal systemic autoimmune disorder with multi-organ failure, as manifested by immune dysregulation, polyendocrinopathy, enteropathy, X-linked (IPEX) syndrome in humans and genetically equivalent Foxp3 mutant scurfy mice (3, 4). Due to a complete deficiency in functional Treg, hemizygous male scurfy mice spontaneously develop a lymphoproliferative disease with systemic inflammation, particularly involving the skin, kidneys, lung, and the liver, and resulting in death within four weeks of age (5, 6). The loss of immune homeostasis does not only rely on T-cell dependent mechanisms but is also orchestrated by B cells via production of autoantibodies, including antinuclear antibodies (ANA) (7–11).

ANA represent serological biomarkers of a variety of systemic autoimmune disorders such as connective tissue disease (CTD) and autoimmune liver diseases (AILD) (12). Based on clinical, histopathological, and serological findings, three major immune-mediated liver diseases can be distinguished, i.e., autoimmune hepatitis (AIH), primary biliary cholangitis (PBC), and primary sclerosing cholangitis (PSC). Although most cases match criteria of one of these entities, features of multiple categories may rarely occur concomitantly within the spectrum of AILD, a clinical phenotype referred to as “overlap syndrome” (13). Despite well-defined diagnostic parameters, the current therapeutic armamentarium is predominantly limited to non-specific immunosuppression and/or anti-cholestatic agents. Hence, if insufficiently treated, chronic inflammation and protracted repair mechanisms in AILD can lead to liver fibrosis and ultimately to cirrhosis, the strongest predisposing factor for hepatocellular carcinoma (14, 15). Of note, liver dysfunction was also reported in patients with IPEX syndrome, which is mostly fatal within the first two years of life (16).

The precise pathogenesis underlying AILD and the specific role of Treg in these diseases are still elusive. Thus, the present study sought to characterize the hepatic disease spontaneously evolving in scurfy mice. In this context, we screened sera of scurfy mice for the existence of ANA by an indirect immunofluorescence assay (IFA), analyzed the staining patterns, and tested for autoantibodies against targets specific to or associated with AILD, i.e. antimitochondrial antibodies with three major epitopes (AMA-MIT3), valosin-containing protein/p97 (VCP), glycoprotein-210 (gp210), Kelch-like protein (KL), hexokinase (HK), lamin B1, liver cytosol type 1 (LC1), soluble liver antigen (SLA), liver kidney microsome (LKM), soluble protein 100 kDa (sp100), early endosomal antigen 1 (EEA1), Ge-1, glycine-trytophan protein of 182 kDa (GW-182), and argonaute protein (Ago2). We further examined histopathological alterations, fibrosis-related transcripts, and the cellular components of the inflammatory infiltrates in scurfy liver. Our findings indicate that Treg-deficient scurfy mice harbor clinical and serological features of AILD with overlapping characteristics of AIH and PBC.

AILD are immune-mediated chronic inflammatory disorders characterized by an abrogation of peripheral tolerance against hepatocytes and biliary epithelium (36). Screening for disease-related autoantibodies is essential to facilitate the diagnosis of AIH and PBC, whilst being of relatively minor importance in PSC. Although AIH is widely considered a T-cell mediated disease and the pathogenic role of autoantibodies remains ill-defined, several autoantibodies have been linked to different clinical phenotypes. As such, type 1 AIH (AIH-1, classic type) is associated with ANA and anti-SMA, whereas the rarer, yet more aggressive forms AIH-2 and AIH-3 are defined by the existence of anti-LKM/anti-LC1 and anti-SLA autoantibodies, respectively (12, 37–39). Considering the severe course of disease in scurfy mice, it is important to note that a significantly increased production of anti-LKM, anti-LC1, and anti-SLA autoantibodies was detected in sera of scurfy mice, along with ANA ( Figures 1 , 2 ).

AMA, a serological hallmark of PBC, were by far the most prevalent AILD-associated autoantibodies in scurfy mice (in 82.6%, Figure 2 ), supporting previous findings (31). This is in accord with the predominant cholangitis in scurfy liver ( Figure 5 ), while interface hepatitis and necroinflammatory activity in hepatic parenchyma, as seen in AIH, were observed to a lesser extent ( Figure 5 ). Scurfy mice also revealed significantly elevated levels of anti-gp210 autoantibodies, which are reported to be associated with worse prognosis and higher risk at hepatic failure in PBC (40, 41). More recent evidence suggests that patients with anti-gp210 antibodies were more likely to develop interface hepatitis and lobular inflammation, akin to that in scurfy mice and PBC-AIH overlap syndrome (20, 31, 40). This concurs well with a study demonstrating that the prevalence of anti-gp210 autoantibodies was significantly higher in PBC-AIH overlap syndrome than in PBC and AIH (42). Furthermore, two recently identified biomarkers for PBC, anti-KL and anti-HK autoantibodies, were detected in almost half of scurfy sera ( Figure 2 ). To date, little is known about clinicopathological correlations of these novel autoantibodies, although there is increasing evidence that anti-HK1 autoantibodies may be affiliated with poorer prognosis in PBC (43). Since cytoplasmic dot staining was frequently observed in scurfy sera by IFA, we opted for determination of autoantibodies against distinct cytoplasmic domains known as GW (G (glycine) W (tryptophan)-containing) bodies. These comprise unique cytoplasmic foci which exert critical functions in mRNA processing in the microRNA pathways and have been recently proposed as complementary biomarkers for PBC found in 5-10% of affected patients (19, 29, 44–46). In our study, scurfy mice yielded a significant production of autoantibodies against several GW bodies, i.e., GW182/TNRC6, Ago2, and Ge-1 ( Figure 3 ). Interestingly, these structures do not possess a surrounding membrane, which may make them readily targetable by autoantibodies (19), as opposed to most of the other intracellular autoantigens in AILD which might only be released upon hepatocyte damage. By comparison, EEA-1 is part of the endosome/phagosome pathways with an IFA staining pattern that can resemble that of GW bodies, but comparatively little is known about its clinical associations (47).

Overlap syndromes are estimated to account for up to 20% of patients with PBC, the majority displaying features of both PBC and AIH (48, 49). Previous reports indicate that PBC-AIH overlap syndrome differs from isolated forms concerning disease course, prognosis, and therapeutic responses (50). Serologically, PBC-AIH overlap syndrome is generally defined by the presence of key autoantibodies of PBC and/or AIH, i.e. AMA and anti-SMA autoantibodies, and usually higher levels of transaminases and relative therapy resistance according to the Paris criteria (51). Recent studies, however, suggest that the serological profile of PBC-AIH overlap syndrome seems to be much more complex. Not only were other AILD-associated autoantibodies detected in this entity, including those against GW bodies, LKM, SLA, HK, gp210, KL, and sp100, but also distinct disease-specific serological patterns were reported (20). In this regard, anti-dsDNA autoantibodies have generated considerable interest since the concomitant existence of AMA and anti-dsDNA has been proposed to be highly specific for AIH-PBC overlap syndrome (52). This corroborates with our transfer experiments showing that autoreactive CD4⁺ T cells from scurfy mice were sufficient to activate B cells of recipient B6/nude mice to produce anti-dsDNA autoantibodies along with the AILD-associated AC-11/12 fluorescence pattern ( Figure 4 ). Anti-dsDNA autoantibodies are typically regarded as serological biomarkers of systemic lupus erythematosus (SLE) (53). While it is assumed that autoantibodies in PBC-AIH overlap syndrome might target unique dsDNA epitopes, there is also evidence indicating an association between PBC and SLE as these entities overlap in some patients (20, 54). A higher incidence of CTD in AIH was also reported (55). We have previously shown that scurfy mice exhibit features of CTD, including SLE, scleroderma, and mixed connective tissue disease (9, 10, 56). Further studies are required to investigate the role of functional Treg deficiency as a common pathway in patients with concurring AILD and CTD. Given that anti-dsDNA autoantibodies were only detected in one out of six recipient B6/nude mice (16.67%) in the current study, the results should be interpreted with caution and need to be validated by future studies. Nevertheless, the proportion of B6/nude mice with anti-dsDNA production seems to reflect the proportion of autoreactive CD4⁺ T cells transferred from scurfy mice which are able to induce autoantibody production in the recipient B cell population, as anti-dsDNA autoantibodies were previously found to be prevalent in 15% of scurfy mice (10).

In this study, a wide range of AILD-related autoantibodies in sera of scurfy mice was detected ( Figures 2 , 3 ). This finding can be explained by a polyclonal humoral response due to the uncontrolled expansion of CD4⁺ T cells in the absence of Treg, leading to production of more than a single autoantibody. On the other hand, the relationship between PBC and AIH in an overlapping setting is believed to be multilayered, ranging from sequential presentation of both entities, their simultaneous existence to being part of a disease continuum (57). As such, detection of multiple autoantibodies both in general and in the same patient has been reported in PBC-AIH overlap syndrome (20, 52). An important question to resolve for future research is why some autoantibodies (e.g., AMA, anti-VCP, anti-EEA1) were produced in the majority of scurfy mice, whereas others (e.g., anti-GW182, anti-LKM) were relatively rare.

As crucial effector cells of innate immunity, macrophages play an important role in the hepatic microenvironment via polarization to different phenotypes (classically activated M1-type and a spectrum of alternatively activated M2-type macrophages) under pathological conditions such as liver fibrosis, viral hepatitis, and hepatocellular carcinoma (35). The past five years have witnessed a renewed importance of macrophages in AILD as well. M1- and M2-type peribiliary macrophages were shown to be increased in human and murine PSC, while M1-type macrophages have been associated with enhanced Notch signaling and self-renewing phenotypes of hepatic progenitor cells (58, 59). In a concanavalin A (ConA)-induced AIH mouse model, splenectomy and IL-34 were found to drive M2 polarization which suppressed hepatic fibrosis and inflammation (60, 61). Furthermore, Li et al. demonstrated that cholangiocyte-derived exosomal long noncoding RNA H19 promoted M1 polarization and hepatic inflammation in PBC and PSC (62). These results indicate that M1 polarization seems to exacerbate AILD, whereas an M2-type polarization promotes inflammation resolution [reviewed in (35)]. This substantiates our findings in scurfy liver revealing a strong macrophage differentiation towards the M1 phenotype and a decreased M2-type polarization in a TNFα-dominated proinflammatory microenvironment ( Figures 6 , 7 ) Surprisingly, this contrasts with our previous research which showed M2-polarized macrophages and a significant Th2 deviation in the skin of scurfy mice, as partly found in scleroderma (56). An explanation for these rather divergent responses may be that scurfy mice possibly recapitulate different stages of PBC-AIH overlap syndrome and CTD. This finding can also be attributed to potential differences in the cutaneous and hepatic inflammatory milieu. Further studies, including single-cell profiling, are needed to determine the heterogeneity and functional plasticity of macrophages in PBC-AIH overlap syndrome and organ-specific inflammation patterns in scurfy mice. Although a significant upregulation of profibrogenic or ECM remodeling transcripts was observed in scurfy liver, no collagen accumulation was detected ( Figure 6 ). The finding that total hepatic collagen accumulation, as quantified by liver HYP, was not increased at sacrifice in scurfy mice, despite highly elevated fibrosis-related transcript levels, can be explained by their short lifespan, which necessitated their sacrifice at 15 to 25 days of age. Thus, only chronic inflammation of “wounds that do not heal” will lead from a Th1 T cell and M1-type fibrolytic immune cell response to a less inflammatory Th2 T cell and M2-type macrophage dominated response that drives tissue fibrosis (63–66). This would likely have been the case in scurfy mice if they would live longer than four weeks, as would be expected to occur in immune deficient patients (63–66). This is also in line with previous findings that fibrosis phenotypes, e.g., collagen accumulation, are significantly detected, chemically or macroscopically, after four weeks of age in most mouse models for liver fibrosis. In this context, at least four weeks are required for a statistically significant increase of collagen deposition in the carbon tetrachloride (CCL₄)- and thioacetamide (TAA)-induced fibrosis model, which is the most representative mouse model for panlobular fibrosis (23, 67). An even slower fibrosis progression was noted for the biliary fibrotic Mdr2 knockout mice which displays a significant hepatic collagen accumulation after ten weeks of age (23, 68). This observation was also reported in other age-dependent fibrosis development studies using genetically modified mouse strains in which the upward trend of hepatic HYP levels was clearer at four to eight weeks after birth (68). We therefore hypothesize that the early-onset lethal scurfy phenotype might represent a premature model for the full development of fibrosis and/or that their specific Foxp3 defect may favor both active fibrogenesis and fibrolysis, resulting in no significant net fibrosis. This can be supported by previous findings that MMP-9 and MMP-13, for instance, can be both fibrogenic via tissue remodeling and subsequent repair, but also be fibrolytic in other (resolution) settings (63, 69). Our findings further suggest a potential link between macrophages and CD4⁺ T cell-mediated tissue destruction in scurfy mice. In the aforementioned study by Haeberle et al., spontaneous Th2 cytokine secretion of skin infiltrating CD4⁺ T cells was associated with M2-polarized macrophages in the skin (56). Concordantly, scurfy mice were found to exhibit cartilage degradation and nonerosive arthritis in the paws, comprising CD3⁺ T lymphocytes, B cells, but also neutrophiles and macrophages (9). As previously demonstrated, in liver tissue of scurfy mice, CD4⁺ T cells were predominantly accumulated in periportal areas, while cytotoxic CD8⁺ T cells were concentrated around the bile ducts (31). The same study also identified increased expression of hepatic genes encoding cytokines such as IL-12, which is mainly produced by antigen-presenting cells including dendritic cells, monocytes, and macrophages (31, 70).

Previous reports indicate that, compared to isolated PBC, PBC-AIH overlap syndrome leads to significantly higher rates of unfavorable outcomes including esophageal varices, gastrointestinal bleeding, ascites, need for liver transplant, as well as abatement of 5-year survival (50, 71). Thus, there is an unmet need for therapeutic options which specifically target key molecules in the pathogenesis of AILD. Notably, a growing body of literature demonstrates safety and efficacy of adoptive Treg therapy in a variety of immune-mediated diseases (72, 73). However, available data on the precise role of Treg in AILD have been sparse and in part contradictory. In AIH patients, recent reports have suggested that Treg are fully functional and are not reduced in frequency (74, 75), despite initial studies indicating otherwise [reviewed in (73)]. Some researchers even found increased intrahepatic and peripheral Treg frequency in AIH, which was more prominent in pediatric patients than adults, implicating that intrahepatic Treg might be functionally defective or insufficient for disease control (75–77). Of note, a 4-year-old patient with IPEX syndrome was reported to develop AIH-2 with anti-LKM1 autoantibodies, which points to involvement of Treg in pathogenesis of AIH (78). In contrast, PBC has been clearly shown to be associated with functionally and numerically impaired Treg both in peripheral blood and liver of patients and mouse models (79–81). Further studies have highlighted the significance of Treg in PBC by demonstrating that deficiency in the alpha subunit of the IL-2 receptor (IL2RA; CD25) led to PBC-like liver disease in human und murine setting (82, 83). Importantly, healthy sisters and daughters of PBC patients were found to possess a significantly reduced Treg frequency, indicating a genetic susceptibility to Treg deficiency in PBC (84). In this context, our results might serve as an incentive for future research to determine the functionality and frequency of Treg in PBC-AIH overlap syndrome.

In summary, we confirm and extend previous findings on PBC-like liver disease in scurfy mice and provide further evidence of concomitant aspects typical of AIH. Our results support the hypothesis that Treg deficiency and the consequent breach of humoral tolerance in scurfy mice is key to the spontaneous development of clinical, serological, and immunopathological features of AILD with overlapping characteristics of PBC and AIH. The study sheds a new light on the role of Treg in the pathogenesis of immune-mediated liver disease.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The animal study was approved by Regierungspräsidium Karlsruhe, Germany. The study was conducted in accordance with the local legislation and institutional requirements.

KY: Visualization, Writing – review & editing, Conceptualization, Data curation, Formal Analysis, Investigation, Methodology, Writing – original draft. SH: Conceptualization, Data curation, Formal Analysis, Methodology, Writing – review & editing, Investigation. YK: Visualization, Conceptualization, Data curation, Formal Analysis, Investigation, Methodology, Writing – review & editing. MF: Conceptualization, Data curation, Formal Analysis, Investigation, Methodology, Writing – review & editing, Supervision. S-YW: Visualization, Conceptualization, Data curation, Formal Analysis, Investigation, Methodology, Writing – review & editing. BG: Formal Analysis, Investigation, Methodology, Writing – review & editing. VR: Conceptualization, Data curation, Formal Analysis, Funding acquisition, Investigation, Methodology, Writing – review & editing. KS: Project administration, Conceptualization, Formal Analysis, Funding acquisition, Investigation, Methodology, Resources, Supervision, Writing – review & editing. DS: Formal Analysis, Project administration, Conceptualization, Funding acquisition, Investigation, Methodology, Resources, Supervision, Writing – review & editing. AE: Conceptualization, Funding acquisition, Investigation, Methodology, Resources, Supervision, Writing – review & editing. EH: Supervision, Writing – review & editing, Conceptualization, Formal Analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources.