Women and Primary Biliary Cirrhosis

Abstract

Primary biliary cirrhosis occurs more frequently in women, and previous studies indicated that the average age of primary biliary cirrhosis (PBC) onset makes pregnancy in PBC patients uncommon. However, more recently, improved diagnostic testing has enabled detection of PBC in younger women, including those of childbearing age. This has led investigators to become increasingly interested in the relationship between the ontogeny of PBC and pregnancy. Published cases indicate that the typical age for pregnant women to be diagnosed with PBC is in the early 30s, and that during gestation, pruritus and jaundice are the most common symptoms. During gestation, susceptible women may experience onset of PBC resulting from the drastic changes in female hormones; this would include not only the mitochondrial damage due to accumulation of bile acids but also changes in the immune response during the different stages of pregnancy that might play an important role in the breakdown of self-tolerance. The mechanisms underlying the potential relationship between PBC and pregnancy warrant further investigation. For women first diagnosed with PBC during gestation, or those for whom first appearance of a flare up occurs during and postpartum, investigation of the immune response throughout gestation could provide new avenues for immunologic therapeutic intervention and the discovery of new treatment strategies for PBC.

Introduction

Primary biliary cirrhosis (PBC) is an autoimmune disease of the liver characterized, in part, by a strong gender bias. PBC is nearly ten times more prevalent in women, and onset occurs in peri-menopausal age, usually at around 50 years [1]. However, with the recent use of serum anti-mitochondrial (AMA) autoantibodies and liver function tests—including early stage liver biopsy—the age of disease detection has expanded to include women of childbearing age [2–5]. We previously reported that more controls than patients with PBC were nulliparous [6], and there was significant association between gravidity and PBC. As with other autoimmune diseases, the disease is secondary to genetic and environmental interactions [7–9].

Gender has important implications for autoimmunity [10–15]. Pregnancy has an important impact on some autoimmune diseases, with varying outcomes (remission or exacerbation) [16–21]. The role of pregnancy in the occurrence and progression of PBC remains unclear. It has been speculated that PBC is a preexisting condition that is detected during prenatal medical supervision or that nonspecific symptoms such as pruritus are first noticed during pregnancy [1]. However, studies suggest that pregnancy does indeed create a susceptible environment for the ontogeny of PBC under certain genetic and environmental conditions. In this review, we focus on pregnancy in patients with PBC, using a systematic analysis of the literature. We discuss the possible mechanisms for pregnancy implications in PBC, including the effect of pregnancy on bile acid (BA) formation and secretion, the role of BAs in the progression of PBC, as well as sex hormones, fetal microchimerism, and the gestational immune response. We then describe how understanding the possible role of pregnancy in the etiology of PBC may help to disentangle complex host autoimmune mechanisms, provide new avenues for immunologic reconstitution in patients with PBC, and identify new targets for early PBC intervention therapy.

Pregnancy and PBC

BA Metabolism and Pregnancy

BAs, the major organic solutes in bile, are derived from two pathways. Synthesis of the first type, termed primary BAs, occurs in hepatocytes as a result of oxidation of cholesterol via the key enzyme cytochrome P450 7A1 (CYP7A1). These primary BAs replace daily BA loss via stool. The other, termed secondary BAs, is produced from the hepatocellular basolateral uptake of BAs undergoing enterohepatic circulation in the sinusoidal blood. Following synthesis, primary BAs are conjugated with taurine or glycine, then cross the canalicular membrane to be secreted in bile [22]. In the terminal ileum and colon, following the bacterial modification, the primary BAs form secondary BAs. Ursodeoxycholic acid (UDCA), a widely used treatment for cholesterol cholelithiasis and PBC, is one of the secondary BAs and is formed as a metabolic by-product of intestinal bacteria [23].

It is widely recognized that hormones play a key role in the regulation of various metabolic processes [24]. Studies have shown that sex hormones are responsible for the differences between males and females with respect to BA metabolism [25, 26], and animal studies have provided evidence that estradiol can contribute to cholestasis [27]. During pregnancy, increased estrogen levels can result in decreased bile uptake. Hepatic estrogen conjugates inhibit bile salt uptake via the sodium-dependent bile salt transporter—Na⁺/taurocholate co-transporter (NTCP)—and sodium-independent bile salt transporters—organic anion transporting proteins (OATP). These specific transport proteins are expressed at the basolateral hepatocyte membrane and contribute to the uptake of BAs [28]. Increasing estrogen can decrease the fluidity of the basolateral membrane [29–31], which causes decreased activity of the Na⁺/K⁺-ATPase, further hindering the sodium-dependent BA uptake into the hepatocyte.

In order to keep the balance between synthesis, uptake, and excretion of BAs, the expression of hepatobiliary transporters is tightly regulated by nuclear receptors (NRs), which provide a network of negative feedback and positive feed-forward mechanisms. During pregnancy, the nuclear BA receptor farnesoid X (FXR), the best defined NRs for BAs, regulates the transcription of the majority of genes involved in the negative feedback pathway, serving as the primary mediator of hepatic BAs [32]. The estrogen receptor alpha (ERα) participates in transporter downregulation during normal pregnancies, and activation of ERα during pregnancy is important for the repression of hepatobiliary transporters in particular. Aleksunes et al. reported that 17α-ethinylestradiol, an orally bioactive synthetic estrogen used as an oral contraceptive, was associated with downregulation of uptake gene NTCP, OATP1a1, OATP1a4, and canalicular efflux gene for the bile salt export pump (BSEP) [33]. ERα represses the function of FXR in an estradiol-dependent manner, and this inhibition may result in pro-cholestatic gene expression and heightened hepatic BA levels during pregnancy [34, 35].

Animal studies show a decline in expression of uptake and efflux transporters in the livers of pregnant mice. The reduction of transport proteins affects the uptake and excretion of BAs, BA metabolism genes, and hepatic NR genes. Aleksunes et al. demonstrated that this regulation began as early as gestational day 7 and was inversely related to increasing concentrations of circulating 17β-estradiol and progesterone as pregnancy progressed [33]. This led to reduced protection of the liver from cholestatic injury and consequent cholestatic liver disease. The authors proposed that elevated BAs during late pregnancy are probably due to increased mRNA expression of classic BA synthesis enzymes [33]. Moreover, estrogen-glucuronides in the liver are also excreted into bile by the multidrug resistance-associated protein 2 (MRP2), where they trans-inhibit the BSEP, further impacting the export of BAs [36]. These numerous factors indicate that heightened estrogen levels during pregnancy may be a factor in intrahepatic cholestasis.

Effect of BA on Biliary Epithelial Cell (BEC) Proliferation, Apoptosis, and Cytotoxicity (Fig. 1)

Fig. 1.

Estrogen, progesterone, and prolactin interact with BECs and hepatocytes during pregnancy. Pregnancy creates a susceptible environment for the ontogeny of PBC under certain genetic and environmental conditions. During gestation, the drastic elevation in female hormones, estrogen, progesterone, and, prolactin, affect BA metabolism and BEC proliferation resulting in accumulation of BA in hepatocyte and apoptotic cell death in BEC

Impairment of bile formation or excretion processes results in accumulation of BAs in blood and hepatocytes. When the concentration of BAs exceeds the binding capacity of the binding protein located in the cytosol of the hepatocyte, hydrophobic BAs will produce toxic effects in hepatocytes [37] and BECs [38, 39]. Alpini et al. [40] provided evidence in animal studies that accumulated BAs can stimulate biliary proliferation. For example, following BA feeding, rats exhibited an increase in cholangiocyte proliferation, secretin receptor gene expression, and secretin-induced cAMP levels, similar to levels found in animals with bile duct ligation. Other studies have indicated that BAs can interact with cholangiocytes and alter cholangiocyte proliferation and bile secretion [40–42]. BAs enter cholangiocytes using the Na⁺-dependent apical BA transporter (ABAT), and via this mechanism, conjugated hydrophobic BA taurocholate (TC) and taurolithocholate (TLC) could increase cholangiocyte proliferation. Accordingly, Alpini et al. found that rats fed with TC or TLC exhibited an increased number of proliferating cell nuclear antigen (PCNA)-positive cholangiocytes and bile duct proliferation compared to the control rats. Further, [³H]-thymidine incorporation studies demonstrated that cholangiocytes from BA-fed rats showed increased cellular proliferation. The authors suggested that elevated BA levels could stimulate cholangiocyte proliferation and ductal secretion. To evaluate this hypothesis, they isolated small and large cholangiocytes from rats, fed the rats with TC, TLC, or BA control diet for 1 week, and determined PCNA and ABAT expression and BA transport activity. The results indicated that TC and TLC could stimulate proliferation of small and large cholangiocytes associated with protein kinase C-dependent upregulation of ABAT. Contrary to the proliferative effect of taurocholate and taurolithocholate, ursodeoxycholate and tauroursodeoxycholate inhibited the cholangiocyte proliferation [43, 44]. Inhibition of liver cell apoptosis is one of the main processes underlying the protective effect of UDCA, along with protection of cholangiocytes against cytotoxicity of hydrophobic BAs, inhibition of cholangiocyte proliferation, and stimulation of impaired biliary secretion [45]. This inhibitory effect may provide an explanation for the histological and biochemical improvement of PBC following ursodeoxycholate or tauroursodeoxycholate treatment.

In addition to stimulating biliary proliferation, BAs can also induce the apoptosis of BECs. A study by Lamireau et al. [39] indicated that although conjugated hydrophobic BA TUDC could not induce BEC apoptosis, TCDC, TDC, TLC, and TC could all induce the BEC apoptosis. While TC had only a slight effect on the induction of apoptosis, it had a striking effect on the secretion of monocyte chemotactic protein-1 (MCP-1), one of the potent chemokines that regulate migration and infiltration of monocytes [46], and interleukin (IL)-6, which is critical to triggering autoimmune reactions and contributes substantially to BEC barrier function and wound repair [47].

Specific BAs could exert potent effects on BEC; generally, taurine- or glycine-conjugated BAs seem to prevent its cytotoxic effects, and only unconjugated hydrophobic BAs (chenodeoxycholic acid (CDCA), deoxycholic acid (DCA), lithocholic acid (LCA)) exert cytotoxicity. Benedetti et al. reported that, in vitro, hydrophobic unconjugated BAs induced damage to intracellular organelles of BECs. Their study showed that BECs are susceptible to the cytotoxic effects of unconjugated BAs [48] and that the primary site of damage was mitochondrial [49, 50]. Regarding the enterohepatic circulation of BAs, unconjugated BAs that have undergone de-conjugation by the intestinal bacteria are detectable in the blood circulation [51]. Benedetti et al. [48] further indicated that the mitochondrial alterations induced by DCA and LCA were similar, with a swollen appearance and electron lucent matrix. When BECs were exposed to CDCA, the mitochondria exhibited globular shapes, rarefaction of the matrix, and loss of cristae. As for the damage by BA to the apical membranes of BECs, reports indicate that only the BA lithocholate can cause changes of the BEC apical membrane [48].

The Effect of Pregnancy Hormones on Cholangiocyte Proliferation

Estrogen hormones are essential for the female reproductive system but also play an important role in the control of fundamental functions in other tissues, including the liver. Estrogens are considered immunomodulating hormones [52, 53], probably acting by potentiating the effects of growth factors and should be considered in discussing all autoimmune diseases [54]. There are three major naturally occurring estrogens: estrone (E1), estradiol (E2), and estriol (E3). E2 is the predominant estrogen during reproductive years, both in terms of absolute serum levels and estrogenic activity. During pregnancy, E3 is the predominant circulating estrogen in terms of serum levels, and during peri-menopause, E1 is the predominant circulating estrogen, though its levels drop dramatically in at menopause. Cholangiocytes are the primary targets of damage in PBC, as PBC is characterized by destruction of the BEC lining small intrahepatic bile ducts, and the progressive development of fibrotic chronic liver disease culminating in biliary cirrhosis [55]. The clinical appearance and progression of different cholangiopathies are influenced by the physiology of the female sex and changes in estrogen status in the body [56], so estrogens have been considered to play a pathogenic role in diseases, such as PBC, that preferentially affect the female sex [57].

Cholangiocyte proliferation, which acts as a repair and compensatory mechanism in the bile duct, may influence the outcome of the disease and its evolution toward the terminal ductopenic stage [58, 59]. Estrogens have been considered for many years to play a role in the development and progression of pathologies involving the biliary tree [56]. They may therefore play a role in modulating the growth of cholangiocarcinoma [60] and could modulate the disease progression in PBC [57, 61–65]. Estrogens can modulate cell growth and proliferation in target tissues expressing ERs [66–69]. It has been shown through animal studies and analysis of human liver tissue that estrogens and their receptors may influence the pathophysiology of cholangiocytes, which is characterized by cholangiocyte injury and proliferation [70–73]. Once bound by estrogen, the ER undergoes conformational changes that allow the receptor to bind DNA elements in target gene promoters and activate transcription [66–69]. In this way, estrogens work through interaction with the ERs to induce transcriptional regulation. Most cells of the human immune system express ERs, including CD4⁺ and CD8⁺ T lymphocytes, B lymphocytes, natural killer (NK) cells, dendritic cells (DCs), and macrophages. This suggests that estrogens may be able to modulate the function of these immune cells [74, 75]. ERs have two different forms, ERα and ERβ, that may play different roles in gene regulation. Paech et al. [76] found that, when complexed with the natural hormone estradiol, the two forms produced opposite effects. When bound to ERα, 17β-estradiol activated transcription, whereas interaction with ERβ led 17β-estradiol to inhibit transcription.

Cholangiocytes are the primary cells targeted by immunopathology in PBC. During PBC progression, cholangiocytes will undergo proliferation as well as loss. Alvaro et al. [70] reported that ERs are not found in cholangiocytes of normal liver but are significantly positive in cholangiocytes from patients with PBC. They demonstrated that in PBC, the presence of ERα increased from stage I to stage III, but in stage IV, ERα was absent in association with maximal ductopenia. ERα positivity in cholangiocytes of patients with PBC was markedly lower than primary sclerosing cholangitis and alcoholic cirrhosis. The low expression of ERα in PBC and its disappearance in the advanced histological stages suggests that an estrogenic deficiency could favor the evolution of this disease toward ductopenia. These reports indicate that estrogens, in conjunction with various growth factors, may play an important role in sustaining cholangiocyte proliferation and in depressing cholangiocyte apoptosis [77–79]. However, some results provide evidence contrary to this hypothesis. Floreani [57] reported that dehydroepiandrosterone sulfate, a metabolite of dehydroepiandrosterone/prasterone (DHEA-S), serum levels were significantly higher in PBC subjects and were higher in precirrhotic than in cirrhotic patients. Alvaro et al. [70] found the serum levels of 17β-estradiol in PBC stage IV were significantly higher than PBC stage I. Alvaro et al. suggested that this discrepancy may be a consequence of cholestasis and liver failure. Despite these observations, it seems that endogenous estrogens could contribute to the PBC progression. The above studies measured serum levels of estrogen rather than the presence of the ERs. Granulomas are also a feature of PBC but the mechanisms involved are still unclear [80]. Clearly, more work is needed to resolve these findings.

Parikh-Patel [6] showed a higher frequency of past oral contraceptive use in patients with PBC than controls, though the difference was not significant. Previous studies from our group [81] have indicated that the history of oral contraceptives may not affect the onset of PBC but have shown significant differences between PBC and controls using hormone replacement therapy, regardless of past or current use of oral contraceptives. However, Corpechot et al. [82] considered the use of oral contraceptives to be a putative protective factor instead of a risk factor. To date, no studies have been able to clearly define the link between the use of oral contraceptives or hormone replacement therapy and PBC.

Progesterone, a steroid hormone synthesized by the corpus luteum of the ovaries and adrenal glands during pregnancy, as well as the central and peripheral nervous system may also play a role in cholangiocyte proliferation [83, 84]. A study by Glaser et al. [85] reported a significant downregulation of progesterone secretion in cholangiocytes isolated from rats with experimental bile duct ligation (BDL). Rat cholangiocytes express both nuclear and membrane progesterone receptors, and while chronic administration of progesterone to normal rats appears to stimulate biliary proliferation, this proliferation can be partially prevented by the administration of a neutralizing anti-progesterone antibody. Cholangiocytes from both female and male rats possess the enzymatic pathway for progesterone steroidogenesis and secretion, suggesting that progesterone may be an important regulator of cholangiocyte proliferation.

Prolactin is another hormone responsible for regulating the growth of female cholangiocytes. A study by Taffetani et al. [86] indicated that prolactin stimulates normal cholangiocyte growth by an autocrine mechanism. Normal female rat cholangiocytes express prolactin receptors, whose expression increased following BDL. They reported that long receptor isoform is predominant in cholangiocytes but not in hepatocytes, indicating that the actions of prolactin on the liver are cell type-specific. The administration of prolactin to normal female rats increased cholangiocyte proliferation, while administration of an anti-prolactin antibody to BDL female rats led to decreased cholangiocyte proliferation [86]. Thus, altered prolactin production may affect disease progression, and regulation of prolactin secretion may be helpful for the management of cholangiopathies affecting female patients.

Cholangiocyte proliferation is a stepwise process involving liver fibrosis during acute and chronic cholestasis and finally leading to periportal fibrosis and eventually biliary cirrhosis [87]. Steroid hormone stimulation of cholangiocyte proliferation might contribute to the crosstalk occurring during cholestasis [85, 88]. In summary, these combined studies suggest that proliferating cholangiocytes may have a role in the induction of fibrosis, either directly or indirectly ways [89–92].

Clinical Evidence of Pregnancy-Associated PBC

Cases of Pregnancy with PBC (Table ;1)

Table 1.

Characteristic of 607 pregnancies from 259 patients with PBC

	Case reporta	Retrospective cohort study [2]	Case-control study [4]
Total number of PBC
Patients/pregnancies	41/50	32/50	186/507
Median or mean age (range) (years)	34 (20–47), 33.6±5.3	32 (23–44)	29.5±5.6
Onset of pregnancy
Before diagnosis of PBC	7 (14.0 %)	17 (53.0 %)	499 (98.4 %)
After diagnosis of PBC	43 (86.0 %)	15 (47.0 %)	8 (1.6 %)
Fibrosis stage	34	31	8
0-II		27/31 (87.1 %)	–
I	1/34	–	4/8 (50.0 %)
III	2/34	–	–
I—III	8/34	–	–
II	2/34	–	–
II–III	1/34	–	–
II or III	9/34	–	–
III	6/34	–	1/8 (12.5 %)
III–IV	–	4/3 (12.9 %)
IV	–	–	3/8 (37.5 %)
OS (PBC/AIH)	5/34	–	–
PBC activity during pregnancy		–	–
Stable	27/50 (54.0 %)	20/50 (40.0 %)	–
Flare	20/50 (40.0 %)	30/50 (60.0 %)	–
Improvement	3/50 (6.0 %)	–	–
PBC activity postpartum		–	–
Stable	16/50 (32.0 %)	16/23 (69.6 %)	–
Flare	17/50 (34.0 %)	7/23 (30.4 %)	–
Improvement	2/50 (4.0 %)	–	–
Deterioration	8/50 (16.0 %)	–	–
NA	7/50 (14.0 %)	–	–
Pruritus	18/50 (36.0 %)	16/50 (32.0 %)	15/507 (3.0 %)
Jaundice	14/50 (28.0 %)	–	–
Others—nausea, epistaxis, fatigue, preeclampsia, night blindness, bone pain	4/50 (8.0 %)	–	–
UDCA therapy during pregnancy	–	–	–
Taking UDCA	32 (64.0 %)	–	–
No taking UDCA	2 (4.0 %)	–	–
No information	16 (32.0 %)	–	–
Fetal outcome		–	–
Live	42/50 (84.0 %)	39/50 (78.0 %)	409/507 (80.7 %)
Preterm	7		–
Spontaneous abortion	6 (12.0 %)	20/50 (40 %)	87/507 (17.2 %)
Stillbirth	2 (4.0 %)	1/50 (2.0 %)	–
Voluntary interruptions of pregnancy	0	0	11/507 (2.1 %)
Gender of the fetus	–	–	–
Female	7	–	–
Male	9	–	–
No information	34	–	–

^aRefs. [4, 93–110, 122]

Earlier studies about PBC have been limited to case reports and small cohort studies. It was in 1950 that Ahrens et al. first coined the term PBC [93]. After a detailed study of fertility and PBC, the authors found that PBC was associated with infertility—most of the patients had scanty, irregular, or profuse periods and even progressed to amenorrhoea—but produced no irregularities in patients who did become pregnant. Consistent with these early results, Olsson [94] reported four healthy pregnancies in three patients with PBC; all the pregnancies occurred after diagnosis, the reproductive ages were 22, 25, 33, and 36, and all pregnancies ended with normal deliveries. In contrast, Whelton [95] reported five patients with PBC, four of whom went to term and had vaginal deliveries, and all of whom presented jaundice in the last trimester. The authors suggested that the endogenous sex hormones associated with pregnancy may have had an exacerbating effect on the preexisting cholestatic liver disease. In 1989, Nir [96] reported a 20-year-old pregnant woman diagnosed with PBC who experienced different clinical symptoms, including not only jaundice but also itching. In 1995, Rabinovitz [97] reported a more severe case in which a PBC patient, diagnosed with PBC during the third trimester of pregnancy, gave birth to a normal healthy child before her disease rapidly deteriorated and she had to be listed for liver transplantation. These reports indicate that there may be a relationship between PBC and pregnancy.

Despite reports from these and other studies, we have few clues as to the relationship between pregnancy and PBC. It should be noted that the concordance of PBC in twins is about 60 % [111].

In 2002, standardized questions were used to compare 182 cases of PBC with 225 age- and sex-matched friend controls to examine the role of reproductive factors in PBC. The results showed a significant association between pregnancy and PBC, where patients with PBC had more pregnancies than controls, and had their first child at an earlier age [6]. In 2013, Trivedi et al. [2] performed a retrospective analysis of women with PBC during pregnancy. They identified 50 pregnancies in 32 women and found that liver biochemistry remained stable in 70 % of patients throughout pregnancy, and 72 % had a flare in biochemical disease activity postpartum. Moreover, only 6 % of patients developed progressive disease after delivery. In 2014, Efe published another retrospective study [5], which indicated that 30 % of the pregnancies with PBC were associated with biochemical flares but no maternal deaths. Seventy percent of patients exhibited clinical improvement or stabilization, and 60 % of pregnant patients showed postpartum flare. One patient was referred for liver transplantation after delivery. These studies suggest that female hormones during pregnancy likely drive the immunological changes behind, and contribute to, PBC flare and progression.

In pregnant women, intrahepatic cholestasis of pregnancy (ICP), a pregnancy-related liver disease presenting similar clinical features and onset time during gestation to PBC [112], is the most frequent cause of cholestasis and may therefore warrant further investigation [113]. Like PBC, ICP is characterized by pruritus, elevated alkaline phosphatase (AP), and an increase in the levels of BAs. At this point, UDCA is the most effective therapy. However, whether there is a correlation between ICP and PBC has not been established. Ropponen et al. [114] analyzed the intrahepatic cholestasis of pregnancy as an indicator of liver biliary diseases; when analyzed separately for PBC alone, the rate ratio did not reach statistical significance, but need to rule out the possibility of the small number of occurrences [115]. Otherwise, during normal pregnancy, it has been demonstrated that serum alkaline phosphatase levels can increase significantly in the third trimester [116]. Serum alkaline phosphatase elevation is one of the most important biochemical indicators of cholestasis and PBC [117], but during pregnancy, this elevated serum alkaline phosphatase may be due to production by the placenta [118]. Pregnancy may contribute to the onset of cholestasis in susceptible individuals under certain genetic and environmental conditions, leading to the detection of hepatic diseases during pregnancy.

Case Studies of PBC in Different Female Life Cycles

Based on the observations above, we hope to elucidate whether the changes in female hormones during pregnancy can drive the susceptible individual’s onset and progression of PBC. We therefore consider other phases of the human life cycle that may help to elucidate the effect of female hormones on PBC.

PBC has seldom been reported in childhood, but during pubertal development, peri-menstrual period, pregnancy, and peri-menopausal periods, when females experience hormonal changes, the risk of developing PBC notably increases. Dahlan et al. [119] were the first to report two pediatric-onset AMA-positive patients with PBC: both were overweight girls, diagnosed with PBC at 15 and 16 years of age, respectively. One was reported stable after starting treatment with UDCA, but the other needed liver transplantation. What could explain the early onset and rapid progression of the disease in these patients? The age of puberty appears to be related more to body weight than to chronologic age; heavier minimum weight for height is necessary for the onset and maintenance of regular menstrual cycles in girls over 16 years of age [120], so the patients were experiencing hormonal changes and were consequently at greater risk for developing progressive PBC. Increased awareness of early-onset PBC may create opportunities for preventative measures, including intervention for individuals exhibiting certain risk factors. Following Dahlan’s report, Floreani et al. [121] described a diagnosis of PBC in a 17-year-old girl, the third youngest case of PBC ever reported. This patient appeared healthy but presented with premenstrual itching 1 year prior to diagnosis. Female hormones undergo dramatic fluctuations during menstrual cycle; in order to prepare for pregnancy, estrogen and progesterone hormones are released in the premenstrual phase of the cycle to stimulate the lining of the uterus. If a pregnancy does not occur, the hormone levels decline. The onset of PBC in this patient occurred during a time period that was characterized by rapid female hormonal changes, suggesting that the sex hormone changes accompanying the premenstrual period may have contributed to her PBC onset. Shah et al. [122] reported a case in which a 36-year-old pregnant patient with PBC who stayed off of UDCA after pregnancy exhibited normal liver function during pregnancy but sharp deterioration postpartum. This sharp deterioration may have been the result of the hormonal changes occurring postpartum. And as we know, PBC onset usually occurs in peri-menopausal age at around 50 years [1], and based on the data discussed, sex hormone changes in different female period of women may contribute to an increased risk of PBC onset or progression in susceptible individuals.

Fetal Microchimerism and PBC

The incidence of autoimmune disease varies relative to the onset of reproductive function. Multiparous women may experience significantly altered susceptibility to PBC, as prevalence of PBC has been associated with greater number of pregnancies [6], while nulliparous women have been negatively associated with PBC [81]. These associations suggest that PBC may be a pregnancy-related liver disease, but how pregnancies affect PBC has yet to be determined. Pregnancy has both short-term effects and long-term consequences on the female body; in addition to the physiological changes that accompany pregnancy, fetal cells traffic into the maternal circulation, where they can persist in blood and tissues for decades [123]. The presence of small numbers of allogeneic fetal cells in the maternal system is termed fetal microchimerism, and these fetal invading cells may be the cause of the pregnancy-associated differences observed in patients with autoimmune diseases. It would be advantageous to the study of autoimmune diseases if we could elucidate how the woman’s immune system responds to these “foreign” cells. In 1996, Nelson et al. proposed that fetal microchimerism might, in part, explain the female predilection to autoimmune disease that women retain following pregnancy [124]. In a follow-on study in 1998, it was discovered elevated levels of fetal microchimerism in the blood of women with scleroderma compared to healthy women [125]. This was the first study to explore the radical concept that many autoimmune diseases may be caused by microchimerism, and subsequent studies have extended this line of inquiry to investigate the potential role of microchimerism in PBC. Kobayashi [126] reported that maternal microchimerism is present within the livers of patients with progressive postnatal type biliary atresia, suggesting that biliary atresia could be a graft-vs-host disease (GVHD) triggered by maternal microchimerism that is masquerading as an autoimmune reaction.

PBC shares important features of autoimmunity with chronic GVHD [127, 128]. Clinical onset in both diseases is characterized by a significant rise in alkaline phosphatase, bilirubin, hypergammaglobulinemia, and autoantibodies including anti-mitochondrial antibodies, with only mild elevation in transaminases. Both are associated with dry gland syndrome and other co-occurring autoimmune diseases. The histologic characteristics include mononuclear cell infiltrates into the portal triad with bile duct injury, bile duct loss, and proliferation of bile ductules. Primary biliary cirrhosis is associated with human leukocyte antigen (HLA) class II genotypes, including HLA-DR molecules. Nelson [124] has suggested that fetal microchimerism might be involved in the etiopathogenesis of some autoimmune diseases, including PBC, probably through initiation of a GVHD-like response [129]. This work is of particular interest for PBC because fetal cells may contribute to alteration of the host response to normal environmental pathogens, during pregnancy and in the following years. Such fetal cells could induce loss of tolerance to fetal antigens that allows the maternal immune system to react to microchimeric fetal cells, and thus, fetal hepatocytes may act as immune targets or as a source of foreign antigen leading to chronic inflammatory diseases like PBC. However, Invernizzi et al. found that in addition to fetal cells, fragments of fetal DNA are also present in maternal circulation. These data do not support the hypothesis that fetal microchimerism plays a significant role in the onset or progression of PBC. The peripheral blood of the PBC women did not contain a higher frequency of male DNA than that of the healthy controls, and the presence of male DNA in patients with PBC was not associated with any particular characteristics of the disease [130]. Furthermore, though some studies reported fetal cells and DNA in liver samples from patients with PBC, most also found fetal microchimerism in control livers as well, offering no clear indication that fetal cells present in the liver contribute to the disease [131–135]. Thus, previous studies suggest that analysis of the biological effects of fetal microchimerism in maternal liver requires further study. More recently, workers have emphasized the role of the X-chromosome and also epigenetic alterations [136, 137].

Immune Response in Pregnancy and PBC (Fig. 2)

Fig. 2.

Maintaining immune tolerance during pregnancy. Fine-tuning maternal innate and adaptive immune response occurs to maintain fetal tolerance, which may create an environment for the ontogeny of PBC in susceptible women

Unlike organ transplants, foreign genes donated by the father during pregnancy can be tolerated by the female immune system. Consequently, the innate immune system might play an important role in pregnancy progression. NK cells recognize the major histocompatibility complex (MHC) class I chain-related gene B (MICB), which is a membrane-bound glycoprotein involved in both innate and adaptive immunity through its interaction with natural killer group 2D (NKG2D) receptor. However, during pregnancy, MHC class I expression is reduced on the syncytiotrophoblast and instead express HLA-G. This inhibits NK cells from binding to activating their killer cell immunoglobulin-like receptors (KIR) [138], thereby blocking the cytolytic activity of NK cells and promoting immune regulation [139]. In addition to the low MHC expression, Thellin and Heinen found that during pregnancy, the transporters that have been previously associated with PBC-specific antigen processing were poorly functioning [140]. Interestingly, E2 was found to suppress MICB mRNA as well as surface protein levels in a dose-dependent manner. Estrogens also appear to have an influence on NK cell number [141]. Studies have reported that the number of peripheral NK cells decreases with an increase in estrogen concentration during pregnancy [142–144]. Studies have additionally reported heightened levels of T helper 2 (Th2) cytokines and with a decline in T helper 1 (Th1) cytokines during pregnancy [145, 146], which further decreases NK cell numbers and NK cytotoxic activity thereby reducing the opportunities for loss of immune tolerance. This inhibition of NK cytotoxicity will abate approximately 6 months postpartum [147]. The breakdown of immune tolerance to self-PDC-E2 and the adaptive multilineage anti-mitochondrial response are the key steps toward PBC onset and usually appear as an extended clinical prodrome. The mechanism for loss of self-tolerance in PBC is as yet unclear, and direct cytotoxic activity against autologous BEC has yet to be demonstrated. Thus, the arm of the immune system responsible for bile duct destruction remains to be determined.

Toward that goal, NK, natural killer T cell (NKT), and monocytes are all important cells involved in innate immune response, and each of them may be involved in the pathogenesis of PBC. The role of NK cells in the immunopathogenesis of PBC is not yet clear, though our preliminary data suggest that the number and mean frequency of NK cells in blood of patients with PBC is greater than control subjects. Additionally, isolated NK cells from patients with PBC had significantly higher natural cytotoxic ability and higher expression of perforin, and NK cells were recruited to liver more efficiently, potentially leading to higher cytotoxicity and exacerbated hepatic damage [148]. After 9 months of pregnancy, the natural cytotoxic ability of NK cells returns to normal levels [147], but whether this is associated with postpartum flare up has not been determined.

NKT refers to CD1d-restricted T cells, which are of particular interest due to the recent identification of a specific subset of invariant natural killer T cells (iNKT) that link the innate and adaptive immune responses [149]. iNKT cells exert important regulatory functions through their capacity to produce both Th1 and Th2 cytokines. PBC is characterized by Th1-polarized T cell responses [150], implicating a role for NKT cells in the regulation of autoimmune diseases. Activation of iNKT cells is a critical factor in modulating the natural history and acceleration of the disease, as the ligand-activated CD1d-restricted NKT cells are of great importance to PBC initiation and the evolution from subclinical to clinical disease [151–153]. Our group found that transformation of growth factor beta receptor II in a dominant-negative mouse model enabled NKT to attenuate the development of PBC. We have demonstrated that NKT cells can exacerbate liver injury in PBC [153], and that in patients with PBC, the NKT cell number and hepatic CD1 expression were elevated [151, 152]. The proportion of NKT cells was significantly decreased in the liver of patients with early PBC, but increased with advanced PBC, and the proportion of activated Fas ligand (FasL)-positive NKT cells was significantly increased in the livers of patients with advanced compared to early PBC. Activated NKT cells may therefore contribute to the biliary epithelial cell death resulting in the progression of PBC [154]. E2 can increase the number of NKT cells producing interferon (IFN)-γ and the IFN-γ mRNA expression level [155]. Interestingly, in pregnancy, the numbers of iNKT cells in the peripheral blood do not change, and they become less able to produce the type 1 cytokine IFN gamma between the first, second, and third trimester; however, the cells become very activated in the third trimester [156].

Monocytes are another class of key innate immune effector cells that produce cytokines and chemokines upon activation. E2 has a direct role in the modulation of monocyte immune function [157]. Luppi et al. [158] reported that in normal pregnancy, the number of monocytes was stable throughout gestation, but monocytes could produce higher IL-12 and IL-1β, especially in late pregnancy. Comparatively, monocytes from patients with PBC seem to produce higher relative levels of pro-inflammatory cytokines, especially IL-1β, IL-6, IL-8, and tumor necrosis factor (TNF)-α, which are critical to the inflammatory response that may be essential in the breakdown of self-tolerance [159]. However, whether monocytes participate in the flare observed in late PBC pregnancy is still unclear.

Women also tend to have higher levels of immunoglobulins. During pregnancy, asymmetrical IgG molecules increase in serum, and the placenta is capable of releasing factors that can regulate the relative proportion of asymmetrical IgG molecules and induce modifications of the produced antibodies. Asymmetrical IgG molecules are univalent antibodies and therefore act as antigen blockers [160]. In early pregnancy, IgM levels decrease immediately but then rise rapidly following delivery [161]. As we know, in patients with PBC, beyond cholestasis and the presence of mitochondrial antibodies, the common feature is high levels of IgM. Little is known about the role of hyper-IgM in the pathogenesis of PBC, but the elevated levels of serum IgM in patients with PBC seem to be related to high expression of CD40L that plays an important role in modulation of BEC apoptosis in PBC [162].

CD4⁺CD25^high regulatory T cells (Tregs) are a subset of T cells that can regulate the cellular immune response [163]. Some studies have shown an increase in the peripheral CD4⁺CD25⁺ Treg cell pool during pregnancy [164–166], and Tillburgs et al. [167] indicated that the percentages of CD4⁺CD25 ^high T cells in decidua were significantly higher than that in peripheral blood. Mjosberg and colleagues [168] reported that during pregnancy, Tregs contribute to strict regulation of both Th1-like and Th2-like anti-fetal immune reactions, but Th2-like cells can escape the suppression of Tregs. This would allow for increased IL-4 and potential regulation of potentially detrimental IFN-γ production. However, they did not find any evidence for altered Treg numbers or function during pregnancy. Impairment of CD4⁺CD25^high Tregs could also play an important role in the breakdown of self-tolerance [163, 169]. Forgers et al. [170] reported that numbers of CD4⁺CD25^high Tregs were inversely correlated with disease activity of rheumatoid arthritis in the third trimester and postpartum, but how the numbers of CD4⁺CD25^high Treg cells change in the flare up in the third trimester of pregnancy with PBC warrants further study.

Estrogen modulates the cell-mediated immune response by inducing TH2 cytokines to inhibit TH1 responses [171]. Successful implantation depends upon these cytokines to create an optimal environment [172, 173]. Pregnant females are biased toward Th2-type immunity rather than Th1-type immunity; during pregnancy, IL-10, IL-4, etc. (Th2-associated cytokines) increased, while T helper cells (Th1-associated cytokines), like IFN-γ, IL-1, etc., were decreased. The authors also observed reduced numbers of pro-inflammatory cytokines like TNF-γ and IL-12 [145, 146]. In addition to estrogen, progesterone serves as an important effector in establishing Th2 bias in pregnancy. It exerts immunomodulator action by inducing blocking factors expressed on lymphocytes, altering cytokine secretion, and reducing the production of pro-inflammatory cytokines [174]. Progesterone was found to specifically block trophoblast elicit Th1 immunity and inhibit IL-10 production but upregulate tumor growth factor (TGF)-β secretion [175].

Prolactin has been shown to have a role in immunomodulation [176]; during late pregnancy, prolactin will reach its highest levels to enable female mammals produce milk [177]. Matalka [178] reported that in addition to its role in milk production, prolactin can enhance the function of Th1-mediated response by increasing production of IFN-γ, IL-12p70, IL-10, [179] but not TNF-α, in whole blood [178].

T cell-mediated immune response has been considered to play an important role in the pathogenesis of PBC [180, 181]. The destruction of the biliary tract in PBC is thought to be mediated by auto-reactive liver infiltrating CD4⁺ T helper cells [182]. In the majority of patients with PBC, CD4⁺ and CD8⁺T-cells reactive with PDC are present in the peripheral immune system and liver [183], but cytotoxic T lymphocytes (CTL) are thought to be directly involved in the tissue injury in patients with PBC [184]. Kita reported a 10-fold increase in the frequency of auto-reactive CTL in the liver as compared to the blood in patients with PBC [185]. Nagano [186] reported that IFN-γ and IL-5, IL-6, IL-10, IL-12, and IL-15 were all expressed in most liver biopsies of patients with PBC where IFN-γ was highly expressed in contrast to the regulatory cytokine, IL-10, and IL-2 and IL-4 were rarely detected. The authors believed that both Th1 and Th2 cytokines might play a role in the pathogenesis of PBC. For this study, patients with PBC were all selected at an early stage, so these results require further study in later stage samples. Soon after, Itoh [187] and colleagues used a murine model of PBC to clarify the relationship between the by CD4⁺ T cell cytokine profile and the formation of hepatic lesions. They investigated the elevation of IFN-γ mRNA expression at an early phase before the appearance of nonsuppurative destructive cholangitis. IL-10, which can stimulate antibody production and inhibit the function of Th1 cells and developing fibrosis, showed delayed expression.

Th17 cells can selectively produce IL-17, and play a critical role in the induction of inflammation and the pathogenesis of autoimmune diseases [188, 189]. Th17 cells and Tregs share a requirement for TGF-β, high TGF-β concentrations induce Tregs, and terminal differentiation of Th17 cells requires IL-1β and TGF-β [190]. Studies indicate that during normal pregnancy, the population of peripheral blood Tregs cells increases, while the number of circulating Th17 cells decreases [191]. This ration would be reversed in the case of embryonic death, where the Th17 cells are increased and Tregs decreased [192]. In patients with PBC, the circulating Th17 cells were increased and Tregs were decreased [193], and recently, our group found that this Th1/Th17 imbalance was relevant to the pathogenesis of PBC. Whether a flare up during or postpartum in patients with PBC is associated with changes in the ratio of Th1/Th17/Tregs needs further study.

Future Directions

One of the major difficulties in PBC is the latency time between the onset of the loss of tolerance and the clinical presentation. There is increasing evidence, with respect to autoantibodies, that this loss of tolerance occurs for many years before the onset of symptoms. This makes identification of etiological factors more difficult. There are, however, mouse models of PBC and these should be explored in more detail [194–198]. In addition, there is increasing evidence for the role of specific lymphoid subpopulations as effector mechanisms in PBC, but the role of hormones in the modulation of these pathways remains enigmatic [199, 200]. We also note that researchers should take advantage of newer technologies in both genomics and proteinomics to identify effector pathways that may also lead to therapeutic tools [201]. PBC is in the crossroads now in which there is considerable data on basic science but a relative lack of translation to therapies that will benefit patients. We submit that it is time to apply these advances in basic science to patients and begin further efforts in such translation.