Cellular senescence in the cholangiopathies: a driver of immunopathology and a novel therapeutic target

Abstract

The cholangiopathies are a group of liver diseases that affect cholangiocytes, the epithelial cells that line the bile ducts. Biliary atresia (BA), primary biliary cholangitis (PBC), and primary sclerosing cholangitis (PSC) are three cholangiopathies with significant immune-mediated pathogenesis where chronic inflammation and fibrosis lead to obliteration of bile ducts and eventual liver cirrhosis. Cellular senescence is a state of cell cycle arrest in which cells become resistant to apoptosis and profusely secrete a bioactive secretome. Recent evidence indicates that cholangiocyte senescence contributes to the pathogenesis of BA, PBC, and PSC. This review explores the role of cholangiocyte senescence in BA, PBC, and PSC, ascertains how cholangiocyte senescence may promote a senescence-associated immunopathology in these cholangiopathies, and provides the rationale for therapeutically targeting senescence as a treatment option for BA, PBC, and PSC.

Cholangiopathies: shifting the balance from tolerance to tissue damage

Immune tolerance is a state of immune system unresponsiveness in the presence of immunity-inducing stimuli. In a healthy state, the liver is a highly tolerogenic organ with specialized cells such as resident macrophages (Kupffer cells), resident fibroblasts (stellate cells), and liver sinusoidal endothelial cells limiting reactivity to daily exposures to dietary or microorganism-derived foreign antigens [1]. Crosstalk among immune subsets and between the parenchymal and non-parenchymal liver cells is key to determining the balance between liver immunity and tolerance. Immune-mediated inflammatory diseases of the liver are those in which this balance is skewed toward a sustained, inflammatory response caused by aberrant proinflammatory immune cell activation, a loss of immune suppressive capacity, and increased infiltration of immune cells into the liver [2]. In chronic liver diseases, this sustained pro-inflammatory immune response is central to driving the progression of liver pathogenesis through damage inflicted on bystander cells (epithelial, endothelial) and activation of fibrogenic cells, potentially resulting in malignant transformation [3]. In some chronic liver diseases, the insult is known (e.g., viral infections). For others, such as a subset of cholangiopathies, the etiology remains undefined, and the driver of an overactive immune-mediated response is not naturally resolved creating a shift from tolerance to tissue damage.

The cholangiopathies refer to a category of chronic liver diseases that target the epithelial cells lining the bile ducts (cholangiocytes) and can progress to end stage liver disease due to lack of effective therapies. Consequently, it is estimated that 16% of all liver transplants performed in the United States and 11% of all liver transplants performed in the United Kingdom were for cholangiopathies, underscoring the morbidity of this group of diseases [4, 5]. The cholangiopathies are further subcategorized into genetic, neoplastic, acquired, and immune-mediated (Figure 1). While each cholangiopathy is unique in its presentation and progression, they share common signaling pathways, processes, and immune and fibrotic responses [4]. Indeed, overlap exists in the pathology of a subset of cholangiopathies, specifically those that present with an immune-mediated inflammatory response, namely biliary atresia (BA), primary biliary cholangitis (PBC), primary sclerosing cholangitis (PSC), and IgG4-associated cholangitis. Of these, cholangiocyte senescence has been implicated in contributing to disease progression in BA, PBC, and PSC. Thus, this review is centered on the potential role of cholangiocyte senescence as a central mediator of inflammation, immune cell recruitment, and disease outcome exacerbation in immune-mediated cholangiopathies. Further, we discuss how targeting senescent cholangiocytes could hold therapeutic value for the treatment of these cholestatic diseases.

Fig 1. Types of cholangiopathies.

The cholangiopathies are a group of diseases that affect cholangiocytes, epithelial cells that line the bile ducts in the liver. The cholangiopathies are classified based on disease pathogenesis which can be immune mediated, genetic, neoplastic, or acquired. For immune mediated cholangiopathies, cholangiocyte senescence has been shown to be involved in disease progression for Primary Biliary Cholangitis, Primary Sclerosing Cholangitis, and Biliary Atresia. Figures created with BioRender.com.

Cholangiocytes: the epithelial targets of the cholangiopathies

The biliary system can be divided into extrahepatic and intrahepatic bile ducts with the latter further classified, based on decreasing size and location, into hepatic ducts, segmental ducts, area ducts, septal ducts, interlobular bile ducts, and bile ductules [6]. Correspondingly, cholangiocytes are also classified based on their anatomical location with those that line the larger ducts, “large cholangiocytes,” appearing more columnar in shape with large nuclei and more organelles, while those that line the small bile ducts, “small cholangiocytes,” are more cuboidal in shape with a higher nucleus: cytoplasm ratio [7]. While cholangiocytes make up only 3 - 5% of the total liver cell population, they are essential for liver function and health. Cholangiocytes are polarized epithelia that modify ductal bile composition through the expression of multiple transporters, such as sodium glucose transporter 1 (SGLT1), solute carrier family 4 member 2 (SLC4A2, or AE2), a Cl−/HCO3 exchanger responsible for biliary bicarbonate secretion, increased bile flow, and maintenance of the protective biliary “bicarbonate umbrella”, and several aquaporin water channels [6]. Moreover, cholangiocytes are a barrier epithelium, preventing bile leakage, thus protecting the surrounding liver tissue from the detergent properties of bile.

In addition to these important physiological functions, cholangiocytes have a role as immune mediators. Cholangiocytes express multiple innate immune-related pattern recognition receptors (PRRs), which recognize pathogen associated molecular patterns (PAMPS) and damage associated molecular patterns (DAMPS) including toll-like receptors (TLRs), retinoic acid-inducible gene -I-like receptors (RIG-like receptors), and nucleotide-binding and oligomerization domain receptors (NOD-like receptors) [6, 8]. These PRRs initiate intracellular pro-inflammatory signaling cascades leading to the secretion of cytokines, chemokines, proteases, growth factors, and release of extracellular vesicles [9]. Moreover, cholangiocytes express several class 1 and class 2 major histocompatibility complex (MHC) molecules such as human leukocyte antigen (HLA)-A, HLA-B, HLA-C, and HLA-DR [6], and thus can likely function as antigen presenting cells. Taken together, it is likely that cholangiocytes play a crucial role in the recruitment of immune cells to the biliary area (via secretion of pro-inflammatory molecules) and the interaction with immune cells (via expression of MHC molecules and immune receptors) underscoring the importance of cholangiocytes in intercellular communication with both innate and adaptive immune cells.

The reaction of cholangiocytes to PAMPS, DAMPS, and other forms of insult and injury is not uniform; in fact, cholangiocytes are heterogeneous in their response to injury with some cholangiocytes undergoing a proliferative response, known as the ductular reaction, and others initiating pathways involved in cellular senescence where proliferation is curtailed [10]. While this divergent response is critical in facilitating liver repair, recovery, and restoring liver homeostasis, left unchecked, both ductular reaction and cholangiocyte senescence can have detrimental consequences. The ductular reaction in cholestatic liver diseases is important and reviewed elsewhere [11]; however, for the purpose of this article, we will focus only on cholangiocyte senescence.

Cellular Senescence: relevance to cell-cell communication

First described by Hayflick and Moorehead in 1961, the senescent cellular fate refers to a cell permanently arrested in the cell cycle typically triggered by signals related to tissue or cellular damage, cancer development, embryogenesis, or cell lineage determination [12]. Cell cycle arrest can also result from a cumulative effect of a variety of mechanisms including telomere shortening, DNA-damage, oxidative stress, and chronic exposure to inflammatory molecules all culminating in the activation of the p53 pathway and upregulation of cyclin-dependent kinase inhibitors p16^INK4a (p16) and/or p21^CIP1 (p21) [13]. Currently, no single marker is specific for cellular senescence. Accordingly, senescent cells are typically characterized by multiple markers including multiple DNA damage foci, depicted by phosphorylation of histone H2AX (γ-H2AX) positivity, increased lysosomal beta-galactosidase activity (senescence-associated [SA]β-gal), and high expression of p16^INK4a and/or p21^CIP1. Cellular senescence affects the integrity of organelles. Indeed, nuclear membranes are altered due to a laminin B deficiency, and an increase in reactive oxygen species (ROS) generation promotes mitochondrial dysfunction. Additionally, mitochondria can become enlarged and undergo a metabolic switch from fatty acid to glucose use for energy. Moreover, senescence can cause stem, progenitor, or immune cell dysfunctional by impacting their ability to proliferate and develop into fully mature, functional cells [12].

In addition to cell cycle arrest, other hallmarks of cellular senescence (Figure 2) include apoptosis resistance and hypersecretion of proinflammatory and profibrotic molecules that make up the senescence- associated secretory phenotype (SASP). Apoptosis resistance is an acquired feature of senescent cells where pro-apoptotic signaling is suppressed in favor of the upregulation of senescent cell anti-apoptotic pathways (SCAPs). It is thought that a senescent cell develops this sustained viability to overcome or avoid its own self-destruction by the toxic SASP [12]. There is redundancy in the SCAPs that are activated during senescence, but the key molecules of these anti-apoptotic pathways are the BCL-2 family members, including BCL-xL, and the HIF-1a, PI3-kinase, and p21-related pathways [14].

Fig. 2. Enabling factors and hallmarks of cellular senescence.

Cellular senescence can be induced by a combination of many factors including oxidative stress, DNA damage, and a dysregulated immune response, the latter which can contribute to the development of chronic senescence. The original hallmarks of cellular senescence include cell cycle arrest, resistance to apoptosis, and development of a bioactive secretome known as senescence associated secretory phenotype (SASP). As the field of senescence progresses, emerging hallmarks of senescence are continually being defined such as driving a premature aging phenotype in non-senescent cells and surrounding tissues (inflammaging, immunosenescence), expressing a unique profile of surface markers, or surfaceome, and secreting systemic, circulating biomarkers that can be detected in plasma, bile, and urine and used for prognostic purposes.

The SASP is characterized by the profuse secretion of chemokines, cytokines, growth factors, proteases, extracellular matrix factors, and nucleotides. These bioactive molecules can be secreted as soluble factors or packaged in exosomes and serve as a signaling mediators between senescent cells and the immune system. The composition of the SASP can be cell type specific or reflect the nature of the stress or injury. Moreover, the inflammatory signature is highly dynamic and can change over the duration of senescence persistence. This dynamic nature of SASP composition may determine whether the effect of senescence is beneficial or harmful to tissue homeostasis [15].

Cellular senescence can be categorized as either acute or chronic. In acute conditions, such as events occurring during embryogenesis and wound healing, senescence is thought to be beneficial, marking unwanted or damaged cells for removal by the immune system. Here, the production and secretion of SASP molecules, such as interleukins and interferons, recruit and/or activate immune cells, such as macrophages, natural killer (NK) cells, and T lymphocytes, to promote senescent cell removal [16]. It is thought that NK and CD8+ T cells act first, damaging senescent cells by secretion of granzymes and perforins, with macrophages clearing the damaged cells through efferocytosis, or phagocytosis of dead cells, thus promoting resolution of the inflammatory response. It has been proposed that “helper” senescent cells, or cells that withdraw from the cell cycle but don’t exhibit the SASP also exist. These cells are thought to be involved in promoting stem and progenitor cell fates, pushing them to develop into the appropriate cell lineage [12]. It is purported that helper senescent cells are also necessary for tissue remodeling during development or removal of damaged tissue during the initial phase of wound repair and cancer defense. A study looking at the role of senescent hepatic stellate cells in a mouse model of carbon tetrachloride (CCL₄)-induced liver injury found that depletion of p53, a central mediator of senescence pathways, or the cell cycle inhibitor, p16, worsened liver fibrosis [17]. Here, the authors concluded that cellular senescence played a role in limiting fibrosis in this model of liver injury. In a separate study, using the high-fat, high-cholesterol (HFHC) diet - induced murine model of nonalcoholic steatohepatitis (NASH) followed by a recovery period on normal chow, it was found that a subset of CD8+ T cells with a resident memory phenotype (CD8+ Trm) were key in liver fibrosis resolution [18]. Depletion of this subset of T lymphocytes resulted in increased abundance of hepatic stellate cells in the NASH model supporting a role for the CD8+ Trm cells in limiting fibrosis by inducing apoptosis in fibroblasts. While this latter study didn’t directly explore a role of fibroblast senescence in limiting fibrosis, together these studies suggest that senescence is beneficial in preventing fibrosis potentially through the recruitment of T lymphocytes which can target senescent cells for elimination. Furthermore, this implicates the ability of senescent cells to work synergistically with the immune system to limit tissue injury and promote resolution and restoration of tissue homeostasis.

When cellular senescence persists, due to an intrinsic, anti-apoptotic phenotype or an inability of the immune system to effectively target and remove senescent cells, a chronic senescent state develops.

During chronic senescence, the surrounding tissue is susceptible to SASP-related damage resulting in a sustained inflammatory and fibrotic response. Here, the destructive SASP not only sustains an inflammatory response, but it can also activate a senescent phenotype in surrounding, non-senescent cells [3, 19]. In the liver, chronic cellular senescence, best characterized in cholangiocytes and hepatocytes, impairs liver function, such as abnormal lipid metabolism [15], cell viability and tissue regeneration [20] and can lead to liver fibrosis and cirrhosis [21]. The persistence and pervasiveness of senescence can be explained by what has been called the threshold effect. Simply stated, the threshold effect is where the spread of senescent cells exceeds the capacity of the immune system to clear them resulting in the rapid accumulation of senescent cells in otherwise healthy individuals. There are several reasons why senescent cells could exceed the threshold level including: (i) a chronic insult that is not resolved; (ii) senescent cell expression of markers that allow immune detection evasion; (iii) an immune system failure resulting in the ineffective/inefficient clearance of senescent cells; and (iv) immune response overcompensation from proinflammatory to a sustained immunosuppressive state.

Senescent cells can upregulate immune evasive markers. For example, senescent dermal fibroblasts express HLA-E, a MHCI molecule that binds natural killer group 2A (NKG2A), an immunosuppressive receptor on NK cells. Interaction between HLA-E and NKG2A prevents NK cell release of perforin, granzymes, and other cytotoxic granules [22]. Furthermore, paracrine signaling from the SASP has been implicated in both diminished macrophage phagocytic function and induced phenotypic switching from a proinflammatory to anti-inflammatory phenotype [23]. Thus, senescent cells may avoid immune clearance by promoting overcompensation toward sustained immunosuppression resulting in the inability to eliminate damaged cells [24]. Taken together, senescent cells exhibit adaptations that promote immune cell evasion and/or reduced clearance by immune cells. The ability of senescent cells to impair non-senescent immune cell function, potentially through the SASP, could lead to premature immunosenescence and exacerbation of disease pathogenesis.

Immunosenescence: relevance to aging and disease

In addition to the ability of senescent cells to evade elimination by immune cells, it is possible that senescent cells actively promote immune cell functional decline. Indeed, a dysregulated immune response is a common feature of aging [25]. Here, immune system dysfunction has been implicated in not only predicting unhealthy aging but is proposed as an underlying mechanism driving it. Studies have shown that age-associated alterations in the immune system have detrimental effects on overall health and well-being [26]. These age-associated immune dysfunctions are defined as immunosenescence and inflammaging and affect both the general, innate, and the antigen-specific, adaptive, immune systems. The concept of immunosenescence, first proposed over 40 years ago, is defined by a gradual deterioration of the immune system resulting in impaired function, such as inability to respond to pathogens or elicit lasting immune memory [25]. While immunosenescence is distinct from inflammaging, which is the overactivation of immune responses resulting in a chronic, persistent inflammatory state associated with aging, the two processes are interconnected, orchestrating an imbalance in immune homeostasis leading to disease development. While the processes of immunosenescence and inflammaging are typically associated with morbidities of aging, whether these play a role in senescence-associated cholestatic diseases merits further exploration.

The importance of immunosenescence in driving pathological outcomes, including in the liver, was demonstrated in a recent study in which the DNA repair protein ERCC excision repair 1 (Ercc1), was selectively depleted in mouse hematopoietic stem cells (HSC), the source of all immune cells in the body, causing accumulation of DNA damage and senescence in immune cells. Here, the authors showed that immunosenescence led to impaired immune function and systemic premature aging [21]. It was found that cellular and humoral immune functions were compromised including impaired NK-mediated cell cytotoxicity, and increased expression of PD1+ T cells, a marker of T cell activation and exhaustion. Moreover, a recent study demonstrated that failure in immunosurveillance, the immune cell process of looking for and recognizing pathogens, foreign particles, or infected or damaged cells, resulted in the accumulation of senescent cells. Here, the authors showed that mice deficient in perforin-1, which lack functional NK, NK T cells, and cytotoxic CD8+ T cells, did not effectively eliminate senescent cells resulting in their increased accumulation leading to kidney, liver, and pancreas fibrosis along with other signs of aggravated aging [28].

Whereas these studies demonstrate the detrimental effect of immunosenescence on driving liver tissue pathology, another study suggested that liver disease with chronic immune activation could lead to premature immune exhaustion and immunosenescence [29]. In this small study, the authors looked at T cell subsets among antiretroviral therapy (ART)-naïve, HIV-infected patients with heavy alcohol consumption with or without advanced fibrotic or cirrhotic livers. They found that patients with advanced liver disease exhibited a lower prevalence of CD8+ effector T cells with a higher naïve: memory CD8+ T cell ratio compared to patients without advanced liver disease. While these observations from a small number of patients did not reach statistical significance, the authors state that this decrease in CD8+ effector T cells in patients with advanced liver fibrosis or cirrhosis is suggestive of immunosenescence [29]. Thus, the authors hypothesized that in cases of liver diseases with chronic immune activation, the inability to repair liver damage and restore liver homeostasis could lead to the development of immunosenescence. As we consider the concept of immunosenescence in the context of immune-mediated liver diseases, it is possible that chronic immune stimulation could be driving premature immunosenescence unrelated to aging.

In addition to T cells, macrophages also undergo immune exhaustion [30]. Macrophages play a key role in maintaining liver homeostasis; however, they can undergo age-related impairments. In elderly, it was shown that macrophages have reduced TLR expression and lower expression of MHCII molecules resulting in diminished ability of these macrophages to respond to pathogenic insults and engage in a pro-inflammatory, “M1 ''- like responses; instead, these aged macrophages exhibited increased secretion of immunosuppressive cytokines such as IL-10 [31]. Conversely, other studies have shown that aged macrophages are pro-inflammatory and express decreased IL-10; moreover, this persistently activated subset of macrophages contribute to profibrotic activity through the secretion of TGF-β [32]. These seemingly conflicting accounts highlight the complexity of senescence and the immune response which can be heterogeneous and exist on a continuum instead of an end state.

Macrophages are active participants in immune surveillance and clearance of senescent cells, thus, age-related impairments in their function impacts the effective elimination of senescent cells. In pathological conditions, immunosenescence has been implicated in the reduced phagocytic capacity of macrophages resulting in these cells being classified as senescent-associated macrophages [23]. One study demonstrated that age-related alterations in the peritoneal microenvironment could drive reduced macrophage phagocytic function as young peritoneal macrophages transplanted into aged peritoneal space showed reduced phagocytic function [30]. In another study, co-culture of macrophages with senescent cells diminishes macrophage function and phagocytic capacity [33]. Thus, it is possible that when functionally competent macrophages are recruited to senescent cells, the constant exposure to the SASP could trigger premature macrophage immunosenescence.

This idea that chronic exposure of immune cells to senescent cell SASP could be driving premature immunosenescence especially holds true when considering the threshold effect, i.e., that senescent cell burden exceeds the immune cell capacity to clear them. Hypothetically, this could result in the expansion of senescence to bystander cells, persistent SASP, and promotion of immunosenescence including in macrophages. However, caution needs to be taken when examining the senescent fate of macrophages. When macrophages take on a proinflammatory phenotype, they undergo cell cycle arrest, lysosomal changes including increased SAβ-gal activity, and p16 expression leading some to propose that macrophages with senescent-like characteristics may just be in another state of polarization [34]. Due to the overlap in phenotypes between a senescent cell and an activated macrophage, it is necessary to examine additional markers of senescence, such as p21 expression or γH2AX accumulation, in addition to functional decline, e.g., diminished phagocytic activity, to determine if a macrophage is undergoing immunosenescence. Thus, the importance of defining the immune signatures associated with immunosenescence and inflammaging is relevant to our understanding of disease progression and therapeutic target identification. Accordingly, many studies have shown changes in immune cells marker expression and function during aging including, among several others, altered macrophage function and a general decrease in overall number of CD8+ T cells (summarized in Table 1). Knowing the immune cell signatures associated with immunosenescence can help determine if chronic diseases, such as the immune-mediated cholangiopathies, also present with signs of immunosenescence.

Table 1.

The immune cell profile and circulating cytokine signature associated with immunosenescence and the cholangiopathies Biliary Atresia, Primary Biliary Cholangitis, and Primary Sclerosing Cholangitis.

Condition	Tissue Immune Cell Profile *also found in blood	Circulating Cytokine Signature	Reference
Immune-focused
Immunosenescence	↓ naïve CD8+ T cells* ↑ memory CD8+ T cells ↑ CD45RA+, CD28−, CD57high, CD27− T cells* ↑ proinflammatory CD4+ T cells ↑ CD56dim NK Cells* ↑ B cells expressing TLR7, TLR9* ↑CD141+ dendritic cells* Diminished neutrophil peroxidase production Impaired basophil and eosinophil activity Altered macrophage function	GDF15, ACTIVIN A, TNFR1, CCL4, FAS, CCL3, TNF-α, IL-6, IL-8, IL-15, OPN, PAI1, PAI2, MPO	Reviewed in Borgoni, S. et al, Ageing Res Rev [26]
Cholangiocyte-focused
Biliary Atresia	↑CD4+ and CD8+ T cells ↑B cells ↑Dendritic cells ↑NKT cells ↑Eosinophils ↑Neutrophils ↑Macrophages	IL-2, INF-γ, IL-4, IL-10, IL-13, IL-18, TNF-α, soluble ICAM1	[38-45]
Primary Biliary Cholangitis	↑ CD4+ and CD8+ T cells ↑Dendritic cells ↑Macrophages ↑NK cells ↑MAIT cells	IL-6, IL-8, IL-10, CCL4 (MIP-1β), CCL19, CCL20, CXCL9, CXCL10, CXCL11, CXCL13, INF-α, INF-γ, CCL26 (Eotaxin-3), IL-12 p70, IL-1β, IL-4, IL-5, IL-13	[51, 64-66]
Primary Sclerosing Cholangitis	↑CD103+, CD69+, CD8+ memory T cells ↑ CD4+ T cells ↑Macrophages ↑MAIT cells ↑Mast cells ↑Neutrophils	IL-6, IL-8, CXCL9, CXCL10, IL-10, CCL4, CCL11 (Eotaxin-1), CCL26	[65, 75-76]

Cellular senescence and the immune response: relevance to cholangiopathies

Biliary Atresia: Is cholangiocyte senescence important?

Biliary atresia (BA) is a cholestatic disease affecting neonates in 1 out of ~10,000 live births and is categorized in two forms: acquired/perinatal (affecting 80% of BA patients) and the rarer embryonic/congenital form [35]. Acquired/perinatal BA presents with progressive inflammation of bile ducts leading to fibro-obliteration of both the intra- and extrahepatic bile ducts. Current treatment options include the Kasai procedure (KP), an operation that involves removing blocked bile ducts and replacing them with a segment of the patient’s small intestine [36]. While the etiology remains obscure, it is thought that viral infections (namely Epstein-Barr virus, cytomegalovirus, human papillomavirus, rotavirus, and reovirus), chronic inflammatory or autoimmune-mediated biliary injury, or abnormalities in bile duct development contribute to pathogenesis. A prominent hypothesis is that in BA, viral infection induces biliary injury which is exacerbated by an aggravated inflammatory or autoimmune response leading to progressive bile duct damage resulting in biliary cirrhosis [37]. However, why viral-induced biliary injury leads to such an exacerbated immune response still needs to be elucidated.

The immune response in BA (summarized in Table 1) appears to be predominantly composed of activated CD4+ and CD8+ T cell accumulation around the portal region accompanied by macrophage infiltration [38]. These lymphocytes are notedly present within the biliary epithelia and are thought to contribute to degradation and loss of biliary epithelial integrity. Further analysis of T cell subpopulations in BA livers indicate that CD4+ and CD8+ T cells may be responding to cholangiocyte “self” proteins as well as viral proteins [39]. It was found that cholangiocytes from BA patients aberrantly expressed MHCII molecules, particularly HLA-DR, although there have been conflicting reports on the association of HLAs with BA pathogenesis [40-42]. In another study, it was found that Tbx21 (T-bet) driven induction of CD4+ T helper 1 (Th1) immune response exacerbated bile duct injury in an experimental rhesus rotavirus (RRV)-infected mouse model of BA. Knockdown of T-bet in this mouse model diminished levels of CD3+, CD4+, and CD8+ T cells although overall obstructive cholangiopathy was not different compared to wild type mice [43]. Here, the authors propose that Th1 cell production of INF-γ leads to the activation of CD8+ T cells and NK cells contributing to bile duct obstruction [43]. The role of INF-y in BA liver damage was also ascertained in a separate study in which the authors propose that INF-y secreted by CD4+ Th1 cells and CD8+ T cells is a key contributor to BA liver pathology [44]. Here, the authors found that T cell expression of PD-1 is crucial to suppress T cell production of INF-y; moreover, the authors concluded that expression of PD-1 ligand, PD-L1 (originating from liver Kupffer cells, hepatocytes, liver sinusoidal endothelial cells, and stellate cells) and PD-1 (on T cells) limits liver damage in the RRV-infected mouse model of BA [44].

Recent work identified heterogeneity among BA patients based on immune cell types and gene expression [38]. In this study, the authors proposed classifying BA into three subtypes based on the predominant immune cell infiltrate. Here, BA subtype 1 was found to be associated with plasma cells, B cells, and conventional dendritic cells; BA subtype 2 was associated with dendritic cells, macrophages (including M1 macrophages), neutrophils, and monocytes; and BA subtype 3 was associated with eosinophils, natural killer T (NKT) cells, and CD4+ memory T cells [38]. Moreover, this study found a strong correlation between IL-4, IL-13, and IL-10 expression with poor prognosis in BA patients [38]. In another study, analysis of BA patient plasma at time of KP and over six months post-procedure demonstrated a progressive increase in circulating cytokines IL-2, INF-y, IL-4, IL-18, and TNF-a with soluble intercellular adhesion molecule 1 (ICAM1) identified as a potential biomarker of disease severity [45].

Although existing evidence support the involvement of viral, toxic, genetic, and immune mediators in driving BA pathogenesis, another explanation for a sustained, exacerbated immune response in BA could be premature cellular senescence. As stated earlier, cellular senescence can occur because of telomere shortening. Telomeres are repetitive DNA sequences that cap chromosome ends and protect them from genomic instability [46]. Telomere shortening due to cell division can induce a state called replicative senescence. While telomere shortening and dysfunction is typically associated with aging, it can also manifest in response to oxidative stress leading to stress-induced cellular senescence [47]. A study looking at telomere length in BA patient liver tissue found more hepatocellular telomere shortening in BA compared to controls, a phenomenon the authors state may happen rapidly after birth [48]. Moreover, decreased telomere length correlated with BA disease severity [49]. Additionally, telomere length was shortened in both liver cells and circulating leukocytes in BA patients indicating the potential for peripheral blood leukocyte telomere length to serve as a non-invasive marker of BA disease severity [49]. Taken together, telomere shortening in liver and leukocytes could indicate premature senescence and immunosenescence in infants with BA.

Given that telomere dysfunction and related cellular senescence can happen regardless of telomere length, it is important to examine other markers of cellular senescence when considering the role of biliary senescence influencing disease outcomes in BA [47]. A study done by Sasaki et al. investigated the presence of cholangiocyte senescence in BA liver tissues at early-stage BA, at the time of KP, and at the time of liver transplant [50]. They found that cholangiocytes in small bile ducts from BA liver tissue expressed a significant increase in senescence markers p16 and p21 compared to control tissue. Interestingly, they also found that cholangiocyte p21 expression was increased in early BA compared to late BA while, conversely, cholangiocyte expression of p16 was more robust in late-stage BA compared to early-stage BA. These findings corroborate other studies that implicate p21 in driving early senescence while p16 is more involved in sustaining the cellular senescence response [50]. While the studies on telomere shortening and expression of p16 and p21 in BA patient tissues support a role for cholangiocyte senescence in BA pathogenesis, more work needs to be done to elucidate the mechanisms by which senescence is initiated and how this cellular phenotype and whether/how this cellular phenotype contributes to disease progression.

Primary Biliary Cholangitis: Autoimmune response driven by senescence-related mitochondrial damage?

Primary biliary cholangitis (PBC) is a female-predominant, chronic, progressive cholestatic liver disease affecting the small intrahepatic bile ducts resulting in non-suppurative destructive cholangitis leading to bile duct loss and eventual cirrhosis and liver failure if left untreated [51]. Current approved treatments are limited primarily to ursodeoxycholic acid (UDCA), which is effective in abrogating disease progression in about two-thirds of patients, with obeticholic acid as a second-line treatment option [52]. While the etiology of PBC is unclear, it is an autoimmune disease characterized by the presence of anti-mitochondrial antibodies (AMA) in 95% of patients and anti-nuclear antibodies (ANA) in about one-third. More specifically, the loss of tolerance to cholangiocytes has been associated with AMA targeting of the E2 subunit of pyruvate dehydrogenase complex (PDC), a mitochondrial enzymatic complex involved in converting pyruvate to acetyl CoA [53]. However, why generation of AMA and ANA happens, and their role in pathogenesis in PBC is still relatively unknown. The potential role microbial agents play in PBC pathogenesis is supported, in part, by the association of PBC with recurrent urinary infections and by studies showing intestinal microbiome dysbiosis in PBC. Work in this area has found sequence similarities between certain bacterial strains and PDC-E2 suggesting that molecular mimicry may promote an immune response against bacterial antigens and aberrant targeting of self PDC-E2 [54]. While evidence exists for the role of molecular mimicry in the generation of AMA, how immune cells are exposed to intracellular PDC-E2 and why only cholangiocyte PDC-E2 (a ubiquitous cellular protein) is targeted is not entirely clear. Although immune system exposure to PDC-E2 is thought to occur during cholangiocyte apoptosis, the contribution of cholangiocyte senescence to this process remains unexplored.

Accumulation of senescent cholangiocytes is associated with PBC. Indeed, expression of senescence markers p21 and p16 is increased in small bile ducts in PBC patient tissue compared to healthy controls. Moreover, cholangiocyte expression of p21 and p16 was found to be more extensive in later stages of PBC (stages 3-4) compared to early stages (stages 1-2) indicating the correlation of cholangiocyte senescence with disease progression [55]. Additionally, it was found that p16 expression in bile ductules significantly correlated to failed patient outcomes after UDCA treatment underscoring the pathological consequences of cholangiocyte senescence both in promoting PBC disease progression and potentially preventing adequate response to pharmacologic intervention [55].

Mitochondrial dysfunction is a prominent characteristic of cellular senescence. During senescence, mitochondrial damage results in release of mitochondrial DNA (mtDNA) that can induce the cGAS-STING pathway, an interferon related signaling pathway that plays a role in cellular senescence and SASP pathways [56, 57]. The interferon pathway has been implicated previously in pathogenesis of PBC [58], and aberrant activation of interferon signaling is linked to immune-related pathologies such as inflammatory autoimmune diseases [59]. Analysis of RNA expression profiles of serum depletion or glycochenodeoxycholic acid (GCDC)-induced cholangiocyte senescence revealed upregulated expression of an interferon signaling molecule, interferon-induced protein with tetratricopeptide repeats 3 (IFIT3) [60]. IFIT3 was found to be highly expressed in p16 and p21-positive senescent cholangiocytes in small bile ducts of PBC patients and significantly correlated with cholangitis activity. Interestingly, IFIT3 is implicated in innate immune signaling via interactions with the mitochondrial antiviral signaling (MAVS) complex on mitochondria [61]. The role of dysfunctional mitochondria in upregulated IFIT3 expression in senescent cholangiocytes is an intriguing possibility the merits further exploration.

The damaging effects of senescent cholangiocytes can be attributed to the sustained, elevated secretion of fibroinflammatory molecules through SASP. Indeed, the SASP is involved in activation of immune cells and their recruitment to the affected peribiliary area in PBC. In particular, expression of CCL2 and CX3CL1, proinflammatory chemoattractants for macrophages, were shown to be robustly expressed in PBC patient small bile ducts compared to healthy controls; notedly, the expression of CCL2 and CX3CL1 was below the level of detection in large bile ducts in PBC patient tissue indicating the specificity of these proinflammatory signaling molecules in modulating the immune response towards small bile ducts [62]. However, knockdown of CX3CL1 and its receptor CX3CR1 in a PBC mouse model did not lead to significant decrease in disease severity [52]. Further, profiling of serum cytokines in PBC patients who did not respond to UDCA treatment found higher levels of CCL19, CCL20, CXCL9, CXCL10, CXCL11, and CXCL13 compared to disease controls and healthy patients (summarized in Table 1); moreover, expression of these cytokines significantly correlated with increased alkaline phosphatase levels [63]. Of these cytokines, it was found that CCL20 is secreted by senescent cholangiocytes and CCL20 levels corresponded to progression of cholangiocyte senescence. Notedly, CCL20 expression can be used to identify high risk PBC patients; thus, further emphasizing the role senescent cholangiocytes and cholangiocyte SASP play in PBC pathology.

Other cytokines that have been implicated in PBC pathogenesis include IL-6, IL-8, IL-10, CCL4, Eotaxin 3 (E3, CCL26), IL12, p70, IL-1β, IL-4, IL-5, IL-13, INF-α and INF-γ, all of which were demonstrated to be highly expressed in PBC patient sera compared to healthy control sera [64-66]. As previously discussed, the interferons signal to T cells, which along with B cells, are the major immune cells known to drive destruction of the biliary epithelia in PBC [67, 68]. Indeed, overexpression of INF-y in the 2-OA-OVA mouse model of PBC promoted heightened liver inflammation with increased infiltration of CD4+ and CD8+ T cells [64]. In addition to CD4+ and CD8+ T cells, other immune cells accumulate in PBC liver tissue including dendritic cells, macrophages, NK cells, and MAIT cells, a subset of T cells that express the invariant alpha chain T cell receptor [51]. While normally protective, chronic activation of these immune cells is thought to contribute to ductopenia in PBC. Interestingly, immunotherapy treatment in PBC patients has largely been disappointing indicating that immune dysregulation in PBC may not be entirely caused by autoreactive immune cells but possibly due to a chronic, unresolved stimulant causing an aberrant immune cell response [63].

Primary Sclerosing Cholangitis: A disease of premature immunosenescence?

PSC is a chronic, cholestatic, immune-mediated disease characterized by progressive fibroinflammation leading to the destruction of the intra- and extra-hepatic bile ducts resulting in biliary strictures, ductopenia, and progression to end-stage liver disease [69]. While the etiology is still obscure, the current understanding is that PSC is influenced by predisposing genetic factors and environmental exposures in addition to gut-liver interactions that can impair biological processes, disturb tissue integrity, and promote aberrant immune responses [4]. Approximately 80% of PSC patients have concurrent inflammatory bowel disease (IBD) with ulcerative colitis being the most prominent form [70]. There is no effective pharmacotherapy for PSC with liver transplantation currently the most viable therapeutic option.

Senescent cholangiocytes are a prominent feature of PSC patient bile ducts [71] and have been implicated in disease pathogenesis. Indeed, senescent cholangiocytes are detected throughout the course of disease and with increasing numbers of senescent cholangiocytes correlating with disease severity [12]. This persistence of senescent cholangiocytes in PSC is destructive. Indeed, studies with cells isolated from PSC patient tissues and animal models of PSC, such as the Mdr2 mouse, have shown that senescent cholangiocytes secrete bioactive molecules, such as IL-8, CCL2, and PAI-1, that attract immune cells, activate a proinflammatory immune cell response, and promote activation and accumulation of fibroblasts [19]. Furthermore, it was found that senescent cholangiocytes could initiate senescence in non-senescent bystander cells in vitro [73].

In patient samples and tissue from animal models of PSC, immune cell and fibroblast accumulation tended to be higher around bile ducts with increased expression of senescent markers. Indeed, when cellular senescence is induced specifically in cholangiocytes in a mouse model of cholangiocyte-specific, conditionally depleted Mdm2, which triggers a p21-driven senescence response, there was increased infiltration of macrophages and progressive accumulation of fibrosis compared to controls indicating that cholangiocyte senescence drives a local fibroinflammatory response in vivo [74]. Moreover, the Mdm2 mouse model showed an increase in hepatocytes expressing markers of senescence suggesting that senescent cholangiocytes drive paracrine-induced senescence in surrounding hepatocytes. Treatment of the cholangiocyte specific Mdm2-depleted mice with diethoxycarbonyl-1,4-dihydro-collidine (DDC) diet showed a worsening of cholestatic liver injury compared to DDC-treated wild type mice. Moreover, in this model, recovery of biliary injury following partial hepatectomy was impaired supporting that senescent cholangiocytes impede liver regeneration [74]. These data establish that senescent cholangiocytes themselves are enough to drive a peribiliary fibroinflammatory response, and when combined with liver injury, exacerbate the injury and impair liver regeneration.

The current understanding of the biliary immune microenvironment in PSC continues to evolve (summarized in Table 1). Analysis from PSC patient sera has demonstrated increased levels of IL-6, IL-8, CXCL9, CXCL10, IL-10, MIP-1β (CCL4), Eotaxin 1 (E1, CCL11) and E3 compared to sera from healthy controls [65]. Interestingly, E1 and E3 are chemoattractant for CCR3-expressing eosinophils, basophils, mast cells, and dendritic cells; furthermore, E3 is a functional ligand for CX3CR1, a receptor found on NK cells and CD8+ T cells [65]. A recent analysis of cells obtained from cytobrush sampling of PSC patient bile ducts showed an enrichment in neutrophils and CD103+ CD69+ CD8+ effector memory T cells, which the authors classified as biliary tissue-resident T cells, compared to non-PSC controls [75]. Additional studies have found the accumulation of macrophages, mast cells, and mucosal associated intraepithelial T-lymphocytes in addition to CD4+ and CD8+ T cells [76]. Additional studies exploring temporal changes to the biliary immune microenvironment over the course of disease will inform on disease progression and, ideally, reveal additional therapeutic opportunities.

While senescent cholangiocytes likely initiate immune cell recruitment to the peribiliary area, whether a dysregulated immune cell response promotes accumulation of senescent cholangiocytes remains an area of investigation. However, the importance of the immune response in PSC is well recognized. In support of this, blocking immune cell recruitment in mouse models of PSC diminishes biliary fibrosis and liver injury [77]. To underscore this, a recent study determined that mast cell accumulation in the peribiliary region of Mdr2 mice modulated levels of biliary senescence through secretion of histamine. Here the authors showed that by knocking out expression of an enzyme involved in histamine synthesis in the Mdr2 mouse, biliary senescence and liver fibrosis was reduced [78]. These findings provided direct evidence that the inflammatory response of an immune cell is enough to drive senescence in other cell types.

Thus, a conceptually attractive model is beginning to emerge regarding cholangiocyte senescence, the immune system, and PSC (Figure 3). In this model, senescence prevents damaged or unwanted cholangiocytes from proliferating, marking them for elimination and, via the SASP, recruits immune cells for their own destruction and prevention of further tissue damage. However, if these senescent cells can’t be eliminated, via an undefined process, the persistent SASP sustains an immune response turning a usually helpful process into one that is harmful. A sustained immune response can, in turn, promote further inflammation, fibrosis, and tissue damage. Moreover, persistent SASP may also induce senescence in neighboring cells thereby promoting premature senescence. If this occurs in the accumulated immune cells, immune cell function may decline, prevent resolution of tissue injury, and skew the balance from resolution to disease progression. Hence, a vicious cycle could emerge in which senescence begets senescence. Targeting the immune response itself is one way to break this cycle; however, although BA, PBC, and PSC patients present as immune-mediated inflammatory disorders, the lack of response to immunosuppressant therapy, particularly in PSC patients, implicates the complexity to the immune response among these cholestatic diseases [70]. While immune-modulating therapies are being considered in the treatment of cholangiopathies, targeting cholangiocyte senescence and SASP may be an additional option to suppress the immune response and limit disease progression. By removing senescent cholangiocytes, a potent and sustained source of proinflammatory molecules, disease severity in BA, PBC, and PSC could be lessened allowing for a shift towards tissue restoration.

Fig. 3. Persistent cholangiocyte senescence and SASP exacerbates disease pathogenesis.

Unresolved cholangiocyte senescence sustains a persistent SASP that can drive senescence in non-senescent bystander cells leading to an exacerbated fibroinflammatory response and potentially inducing premature immunosenescence. While the initial insult is still obscure, the reasons for unresolved cholangiocyte senescence could be due to a dysregulated immune response or senescent cholangiocyte expression of immune evasion markers.

Targeting senescent cells to improve disease outcomes: How might it work?

Senolytics are a group of drugs that selectively induce apoptosis in senescent cells by targeting molecules in the senescent cell anti-apoptosis pathways (SCAPs). Initial senolytic studies used naturally derived products or those in use for other indications such as cancer therapies. These included dasatinib (D), a tyrosine kinase inhibitor which promotes apoptosis by targeting the Src kinase family, quercetin (Q), a flavonoid which targets PI3-kinase delta in addition to other BCL-2 family members, and fisetin, another flavonoid that also targets BCL-2 family members, specifically BCL-xL, in addition to HIF-1α [79, 80]. Since there is heterogeneity among senescent cells and redundancy among the SCAPs, in some cases it is more effective to combine multiple senolytic therapies. For example, neither D nor Q individually promoted apoptosis in senescent mouse embryonic fibroblasts, but the combination of D+Q created a senolytic effect [12]. In addition to D, Q, and fisetin, additional small molecules with senolytic activity include BCL-2 family inhibitors such as Navitoclax (ABT263), ABT-737 and Nutlin3a, Bcl-xL inhibitors such as A1331852, A1155463, and UBX1325, inhibitors of MDM2/p53 interactions such as UBX0101, and mitochondria-targeted tamoxifen (MitoTam) among others [81]. However, some of these first-generation senolytics, such as Navitoclax, promote apoptosis or cellular dysfunction in non-senescent cell types such as platelets and immune cells, resulting in off-target effects. Additionally, treatment with Nutlin3a itself could induce senescence [12].

To overcome limitations associated with some of the first-generation senolytics, second generation strategies have been developed and include galacto-oligosaccharide coated nanoparticles with toxic cargos that are preferentially engulfed by senescent cells [82], proteolysis-targeting chimeras (PROTAC) that both bind target proteins and promote their degradation via the proteosome [83], immunomodulators, vaccines that use senescence-associated T cells as potential targets [84], and chimeric antigen receptor (CAR) T cells [85]. In this latter study, development of CAR T cells selectively and safely targeted and eliminated senescent cells in mouse disease models and reversed pathology and extended life span. Here, CAR T cells were engineered to target plasminogen activator receptor uPAR, a senescent cell - specific antigen discovered through a screen of candidate markers from RNA sequencing datasets generated from three different models of senescence and confirmed to be expressed on the surface of senescent cells. Treating two mouse models of liver injury, CCL₄-induced liver injury and a mouse model of nonalcoholic steatohepatitis, with the uPAR-specific CAR T cells resulted in reduced senescent cell burden, reduced liver fibrosis, and improved liver function [85].

Of the first generation senolytics, dasatinib and the flavonoids quercetin and fisetin were shown to be safe and effective in vivo. Given this, a study was performed that tested a panel of flavonoids for effects on senescence in mouse embryonic fibroblasts induced by oxidative stress [80]. Of this panel that included resveratrol, fisetin, curcumin, and quercetin, it was found that fisetin was most effective in reducing the number of SAβ-gal positive cells with no adverse effects even when given at high doses. The authors found that fisetin was not only effective at targeting mouse embryonic fibroblasts, but that it targeted other senescent cells such as mesenchymal/progenitor T lymphocytes, natural killer cells, and endothelial cells in a progeroid mouse model of advanced aging.

Although promising in some animal models, not all disease models with underlying cellular senescence, or even all tissue types within a disease model, will respond equally to senolytic treatment. In a study looking at the effects of D+Q treatment on cancer development in a mouse model of nonalcoholic fatty liver disease (NAFLD) – induced hepatocellular carcinoma (HCC), the authors found that treatment with D+Q did not remove senescent cells from the mice as determined by expression of p16 and SAβ-gal staining; moreover, treatment of the NAFLD-induced HCC mice with D+Q caused a modest increase in pro-tumorigenic outcomes [86]. These results underscore senescent cell heterogeneity, and the unique roles senescent cells play in pathogenesis of disease.

Nevertheless, due to the safety and efficacy of dasatinib, quercetin, and fisetin, the majority of the clinical trials that are planned, completed, or ongoing use these senolytics. In a first-in-human pilot study of senolytics, short-term, intermittent oral administration of D+Q was given to a small cohort of patients diagnosed with idiopathic pulmonary fibrosis (IPF), a chronic, fibrotic lung disease, over three consecutive days in three consecutive weeks [87]. Physical and pulmonary function tests in addition to clinical chemistries and circulating SASP markers were evaluated before and after treatment with D+Q. It was found that intermittent treatment of IPF patients with D+Q significantly improved physical function (gait speed, chair-stands) one week after completion of the drug regiment although reported respiratory function and fatigue remained unchanged. Circulating SASP markers, such as IL-6, MMP-7, and TIMP, showed a trend toward reduction after D+Q treatment; however, this decrease did not reach statistical significance. Overall, treatment of IPF patients with D+Q did result in clinically meaningful improvements in physical function demonstrating the feasibility and potential impact of the use of senolytics to treat IPF patients. Further studies using D+Q in larger, controlled clinical trials are needed to better evaluate the use of these senolytics in improving disease outcomes in IPF patients. In another small study, D+Q was used to treat patients with diabetic kidney disease (DKD), the most common cause of end-stage kidney failure in which senescent adipocytes have been determined to contribute to worsened disease outcomes [14]. In this study, a single, three-day course of oral D+Q was administered to DKD patients and senescent cell abundance was assayed 11 days after the last dose. It was reported that the single, brief course of D+Q attenuated adipose tissue and skin senescent cell burden, decreased adipose tissue macrophage accumulation, enhanced adipocyte progenitor replicative potential, and reduced key circulating SASP factors such as IL-1a, IL-6, IL-2, MMP-9, MMP-12, IL-1RA, and GM-CSF.

This brief course of D+Q was likely effective since senescent cells take weeks to months to develop and do not divide; thus, eliminating only 30% of senescent cells can be sufficient to be effective and alleviate dysfunction in preclinical studies [14, 87]. Therefore, intermittent administration of senolytics should be sufficient to alter disease outcomes. While treatment with senolytics represents a promising therapy for diseases with underlying senescence, Kirkland and Tchkonia propose special guidelines before starting any clinical trial with senolytics. First, senescent cells should be present in association with the phenotype. Second, individuals without senescent cells should not have the phenotype. Third, inducing accumulation of senescent cells should cause the phenotype. Fourth, clearing these induced senescent cells should alleviate the phenotype. Fifth, clearing naturally occurring senescent cells should alleviate the phenotype. Sixth, the drug should induce few to no side effects. Seventh, administering the senolytic intermittently should be effective. Finally, the candidate drug should alleviate multiple age-related conditions [12].

Using these guidelines, it is proposed that senolytic therapy for patients with BA, PBC, and PSC could be beneficial. First, it was shown in all three conditions that cholangiocyte senescence is an underlying condition and contributes to disease pathogenesis. Secondly, it was shown that treatment of mouse models of PBC and PSC with senolytics alleviated senescent cell burden, biliary fibrosis, and improved disease outcomes. A recent study by Sasaki et al. demonstrated that treatment of their in vitro model of etoposide-induced cholangiocyte senescence with senolytics A-13311852, Navitoclax, Dasatinib, and Dasatinib + Quercetin induced apoptosis and subsequent elimination of senescent cholangiocytes [55]. Here, the authors suggested that senolytic drugs could be useful in treatment of PBC, especially in patients who are non-responders to UDCA. Additional studies looking at the depletion of senescent cells in the Mdr2^−/− mouse, a mouse model of PSC that shows increased expression of senescence markers, p21, p16, and yH2AX in cholangiocytes, found that depleting senescent cells improved fibrosis and inflammation [88-90]. Here, p16 depletion in Mdr2−/− mice via tail vein delivery of a “Vivo-Morpholino” [89], p16-deficient Mdr2 mice (p16^−/−XMdr2^−/−, [88]) or induced, selective depletion of p16 positive cells by crossing the Mdr2^−/− mouse with the INK-ATTAC mouse [91] improved inflammation, as indicated by decreased liver expression of inflammatory markers TNFα, IL-1b, and MCP-1, and diminished fibrosis compared to controls. Similarly, Mdr2^−/− mice treated with senolytic drugs showed a similar reduction in inflammatory markers and fibrosis along with a reduction in the expression of p16 and p21 [88, 90], thus providing preclinical data supporting the use of senolytics in PSC.

Biomarkers and therapeutic targets for cholestatic liver disease: What role can SASP play?

While senescent cholangiocytes accumulate in tissues of patients with BA, PBC, and PSC, it’s likely that not every patient is a candidate for senolytic therapy. Disease progression is different in each patient and determining senescent cell burden in liver tissue requires invasive techniques. Thus, less invasive means of assessing senescent cell burden in patients with immune-mediated cholangiopathies is pertinent. Additionally, more reliable biomarkers are needed to predict outcomes and to serve as surrogate end-points and/or stratifiers for clinical trials. Aging-related molecular pathways have been implicated in chronic cholestatic conditions [92]. Studies have assessed the genomic methylome profile in PBC and PSC patient peripheral blood [93, 94]. In one study, genome wide methylated CpG sites (DMCs) and differentially methylated regions (DMRs) were assessed in patients with PBC or PSC. Although systemic methylation changes were subtle, PBC patients showed DNA methylation profiles significantly associated with the IL-17 signaling pathway, a key pathway in immune regulation. Additionally, it was determined that PSC patients with IBD had DMRs associated with TGF-β signaling among others, and PSC patients without IBD had DMRs associated with leukocyte transendothelial migration. Intriguingly, PSC patients with or without IBD had DMR-associated genes enriched in oxidative stress and senescence pathways [93]. A separate study of PSC patients assessed whole blood DNA methylation (DNAm) using the DNAm signature derived from DMCs associated with the Horvath clock, a proposed epigenetic measure of chronological and biological age that is accelerated in certain pathologies. It was found that age accelerated DNAm patterns were significantly higher in patients with PSC compared to healthy controls and correlated with worsened disease outcomes such as increased serum alkaline phosphatase and alanine aminotransferase, enhanced liver fibrosis, and increased prevalence of cirrhosis [94]. While these studies demonstrate the utility of using epigenetic profiling to find candidate biomarkers, circulating, secreted molecules associated with cellular senescence are also being investigated.

Currently, methods to determine senescent cell burden by detecting circulating SASP markers in patient serum, urine, or bile are being explored. As components of the SASP are profusely secreted by senescent cells, components of SASP can be detected in serum or plasma using readily available assays [95]. Since SASP is a core component of senescent cells, utilizing circulating levels of SASP components as an indicator of senescent cell abundance may be useful in identifying patients who have underlying senescence co-morbidity, help to stratify patients for senolytic clinical trials, and potentially be used as a surrogate biomarker for disease progression. Ultimately, this may also help determine if senolytics could be an appropriate treatment option for individual patients.

However, there is significant overlap between proinflammatory molecules secreted as part of SASP and those that are generally associated with inflammation. Therefore, it becomes imperative that a unique, distinct profile of SASP biomarkers is defined. A study correlating the circulating levels of SASP molecules with advanced age and adverse health events found that of a panel of 24 biologically relevant SASP proteins, 17 were significantly associated with chronological age [96]. Of these, growth factor differentiation factor 15 (GDF15) and ACTIVIN A were the strongest candidate biomarkers followed by TNFR1, CCL4, FAS, CCL3, TNF-α, and IL-6. In addition, GDF15 and osteopontin (OPN) were significantly associated with the frailty index in all participant groups. Interestingly, while these SASP components can be reliably quantified in human plasma, when determining their levels in senescent endothelial cells and preadipocytes, it was found that GDF15 and OPN, along with IL-8, were abundantly secreted by senescent endothelial cells [96, 97] while senescent preadipocytes predominantly secreted ACTIVIN A and IL-6 along with IL-15 and PAI2. Circulating levels of ACTIVIN A is also linked to premature immunosenescence [98], and myeloperoxidase (MPO) and PAI1 have also been linked to systemic cellular senescence [99]. Additionally, SASP molecules have also been detected in extracellular vesicles isolated from patient plasma and urine [100, 101]. A recent study demonstrated that, IL-8, a prominent cholangiocyte SASP component, was elevated in PSC patient serum, and when patients were stratified according to serum IL-8, this potential biomarker predicted transplant-free survival [102]. While the authors did not correlate IL-8 levels to senescent cell burden, this chemokine is an attractive candidate to pursue as a SASP biomarker; further studies using PSC patient plasma to discover novel-senescence-associated biomarkers are actively being pursued.

Just as cellular senescence can be categorized as acute or chronic, SASP can be similarly categorized based on the state of cellular senescence. Early SASP is correlated with increased TGF-β signaling which has been implicated in inducing and maintaining a senescent phenotype [103]. When senescence persists, a late SASP develops which appears to be regulated mainly by transcription factors NF-kB and C/EBPβ with the latter being a key molecule driving the transition from acute to chronic senescence [104]. Several other molecular pathways are involved in regulating SASP either by modulating NF-kB and C/EBPβ expression or by directly inducing gene expression of key SASP molecules such as IL-8 and IL-6. These include p38MAPK, mTOR, JAK/STAT, mitochondrial dysfunction-associated senescence (MiDAS), and PI3k/AKT pathways [81]. Hence, the overall picture of secreted SASP components and the molecular machinery driving its expression is becoming clearer, revealing potential therapeutic targets for many senescence-associated diseases.

Therapies to suppress the SASP are also under investigation. These senomorphic therapies, interventions intended to modulate the senescent phenotype in a cell without killing it, are promising; especially in cells that have limited regenerative capacity in which senolytics would be destructive to tissue integrity. There are two approaches to senomorphic therapies. The first is to target components of the molecular pathways that regulate SASP. These senomorphic drugs include NF-kB inhibitors such as resveratrol, apigenin, kaempferol, metformin, and glucocorticoids, mTOR inhibitors such as rapamycin and everolimus, PI3k/AKT inhibitors such as epigallocatechin gallate, and JAK/STAT inhibitors such as ruxolitinib [81]. The second approach is to target soluble SASP components using neutralizing antibodies such as ABX-IL-8, a humanized monoclonal antibody that binds to IL-8, or Mab-IL-6.8, a monoclonal antibody raised against IL-6 [105]. Of these senomorphic therapies, only metformin, ruxotinilib, and the glucocorticoids, cortisol and corticosterone, are FDA approved drugs. While promising, a limitation to senomorphic therapies is that SASP is not uniform across tissue or cell type and can change based on the senescence inducer; in fact, distinct senescent cell types may uniquely contribute to the circulating SASP signature within a single individual. Moreover, recent data indicates that SASP components are much more extensive and heterogeneous than previously thought as even within a SASP secretome from a single source, soluble SASP can be completely distinct from SASP packaged in extracellular vesicles [106]. Therefore, due to the dynamic nature of SASP, the effectiveness of senomorphic interventions in vivo still need to be elucidated [107].

In addition to detecting circulating SASP makers in patient serum, studies have also shown the utility of using peripheral blood mononuclear cells (PBMC) in predicting immunosenescence. It has been established that the alterations in immunophenotype of lymphocytes in premature immunosenescence shifts from naive T cells to memory T cells that lack expression of co-stimulatory receptor CD28 and increased expression of CD57, a marker of poor proliferative capacity that is also associated with senescence [108]. B cell populations also show a decrease in naive B cells and increase in memory B cells. In an exploratory study looking at the immunophenotype of circulating T lymphocytes in blood samples from testicular cancer survivors (TCS) who were treated with chemotherapy and received at least three cycles of bleomycin, etoposide, and cisplatin, it was found that TCS had decreased proportions of naive T cells and increased memory T cells compared to healthy, age-matched controls [108]. Moreover, further analysis revealed that this circulating subset of T lymphocytes had an increased expression of p16 compared to controls indicating an immunosenescence phenotype in these cells. Whether and how immunosenescence influences cholestatic liver disease progression or could be used as a biomarker of disease remains largely unexplored.

Although SASP is heterogeneous, complex, and dynamic, identifying a comprehensive senescence profile based on previously established “core” circulating SASP components that includes soluble SASP and extracellular vesicle SASP, in addition to circulating lymphocyte composition and lymphocyte expression of senescence markers, may be helpful in establishing a senescence signature unique to each of the immune-mediated inflammatory cholangiopathies that have an underlying senescence co-morbidity. These signatures may prove to be useful for clinical trial patient stratification as biomarkers of senotherapy efficacy can, ultimately, be used in the context of individualized medicine for targeting discrete features of the patient SASP. Based on previously published studies in BA, PBC, and PSC, SASP markers IL-8, IL-6, and INF-y may be potential biomarkers for a cholangiopathy-specific SASP profile although further studies are required to confirm this.

Unanswered Questions & Future Directions

A broader understanding of senescence, the senescence-associated secretome, and how these relate to disease progression of the cholangiopathies is needed. While great progress has been made, several questions still remain. For example, can circulating SASP markers be predictive of underlying senescence, disease progression, and poor outcomes in cholestatic liver diseases? Can these biomarker panels be used to monitor response to therapies, including senolytic and senomorphic treatments? Moreover, the phenomenon of immunosenescence is emerging as a potential pathogenic feature in a variety of diseases; however, whether this form of immune cell dysfunction is involved in the cholestatic liver diseases remains largely unexplored. As the field continues to evolve, and a better understanding of the role of cellular senescence, SASP, and the development of senescence-related immunopathologies in the cholangiopathies emerges, it is critical to reassess how we define and classify the cholangiopathies. In particular, we question whether better ways to label cholestatic liver diseases, in particular, PSC may emerge. For example, could PSC ultimately be relabeled as not only an immune-mediated inflammatory disease, but an inflammatory disease of premature senescence or premature immunosenescence? As the field evolves and if further studies continue to support a prominent role of senescence in the etiopathogenesis of liver disease, it seems possible, if not likely that BA, PBC, and PSC may ultimately be considered diseases of premature senescence of key cell types in the liver, a concept akin to premature aging of the liver. Hence therapeutic advances that slow the aging process, may ultimately prove to be beneficial therapeutic options for BA, PBC, and PSC.

Conclusion

The consequences of cholangiocyte senescence in the immune-mediated inflammatory cholangiopathies is significant. Not only do these cells profusely release a destructive SASP, but they can also drive senescence in non-senescent bystander cells. In diseases such as BA, PBC, and PSC where immune cell infiltration is prominent and located near the bile ducts, these inflammatory and potentially destructive senescent cells are in a prime position to promote a premature immunosenescent phenotype in infiltrating immune cells. The outcome of this may have dire consequences in that immune cells with diminished functional capacity due to premature immunosenescence may not be able to resolve biliary injury or restore homeostasis. Instead, senescence will spread, affecting not only cholangiocytes or immune cells but also additional resident cells including the fibroblasts that accumulate potentially perpetuating a fibrotic response. Utilizing senolytics that have been proven to be safe and effective in small scale clinical trials may be a treatment option to lessen disease severity in patients with immune-mediated cholangitis that also present with indications of cellular senescence as determined by increased serum or urine biomarkers. While targeting senescent cells in the cholangiopathies may not resolve the underlying insult that triggers disease onset, it can help improve disease outcomes. Senescent cholangiocytes are detected in tissues from patients with BA, PBC, or PSC. With the understanding that even a low percentage of senescent cells can be destructive if not eliminated, including senolytic therapy in patients that present with circulating markers of senescence may be beneficial in slowing the progression of inflammation and fibrosis in these liver diseases.