The future of autoimmune liver diseases – Understanding pathogenesis and improving morbidity and mortality

Autoimmune liver diseases (AILD), namely autoimmune hepatitis (AIH), primary biliary cholangitis (PBC) and primary sclerosing cholangitis (PSC), are rare diseases. These days, patients with PBC almost never require liver transplantation. When treated early with ursodeoxycholic acid patients have a normal life expectancy if the disease is diagnosed at an early stage and the patients respond to treatment. Patients with AIH often go into remission with first‐line therapy including corticosteroids alone or in combination with azathioprine. Nevertheless, about one quarter of patients already developed cirrhosis at diagnosis. Those who do not respond to first line standard of care (SOC) have significant liver‐related morbidity and mortality. No approved second‐ or third‐line treatments are available and the drugs are selected based on limited case series and personal experience. Larger trials are needed to develop efficient therapies for difficult‐to‐treat AIH patients. No treatment has been found to alter the natural course of disease in patients with PSC except for liver transplantation. Identifying PSC patients at risk of developing cholangiocarcinoma (CCA) is another unmet need. Current research in all AILD including AIH, PBC and PSC, focuses on improving our understanding of the underlying disease process and identifying new therapeutic targets to decrease morbidity and mortality.


| PRIMARY B ILIARY CHOL ANG ITIS
Primary biliary cholangitis, a non-suppurative destructive cholangitis, is the archetype of an autoimmune disease. Women are affected more often (approximately 9:1) than men and well-defined antibodies targeting the E2 subunit of the pyruvate dehydrogenase complex are a diagnostic hallmark. PBC is associated with a variety of other autoimmune disorders, for example celiac disease, autoimmune thyroid disorders and systemic lupus erythematosus, to name a few. Originally PBC was called primary biliary cirrhosis because patients were often diagnosed in advanced stages of the disease. At present, due to improved techniques, PBC patients are diagnosed in earlier stages with less advanced fibrosis. In the 1980s, PBC was a leading indication for liver transplantation but, nowadays, liver transplantation for PBC is a rare event. Therefore, the name was changed to PBC to mirror the current prognosis of PBC. PBC patients, who are diagnosed at an early stage of the disease and who respond to ursodeoxycholic acid (UDCA) therapy have a life expectancy similar to age-and sex-matched controls. These patients no longer develop end-stage liver disease and the need for liver transplantation is rare. Disease progression is stopped if alkaline phosphatase (ALP), a surrogate marker for inflammation of the bile ducts, normalises or drops by more than 40% within one year after the initiation of UDCA therapy (13-15 mg/kg body weight daily) (Barcelona criteria). 1,2 There are multiple definitions for UDCA treatment failure. UDCA treatment failure occurs in 25%-50% of these patients depending on the criteria. 3 Non-response one year after treatment initiation places these patients at a higher risk of disease progression, including the development of hepatocellular carcinoma and end-stage liver disease. 4 The POISE (Phase 3) trial evaluated obeticholic acid (OCA), a farnesoid-X-receptor (FXR) agonist for the treatment of patients with PBC who did not achieve a reduction in ALP with UDCA to less than 1.67 × upper limit of normal (ULN) or in bilirubin to below 2 × ULN. The primary endpoint was ALP <1.67 × ULN with at least 15% ALP-reduction and normalization of bilirubin at 12 months.
This endpoint was reached both in the 5-10 mg OCA group (with or without UDCA) and the 10 mg OCA (with or without UDCA) group in 46% and 47% respectively. Patients in the placebo group (with or without UDCA) reached the primary endpoint in 10% of cases.
As expected, the major side effect of OCA treatment was dosedependent pruritus, which led to discontinuation of treatment in up to 10% of patients. In addition, an increase in LDL-cholesterol was observed and the long-term cardiovascular risk in PBC patients treated with OCA must still be determined. 5 In 2018, the BEZURSO (phase 3) trial was published, evaluating bezafibrate in combination with UDCA in non-responders. The primary endpoint was complete biochemical remission after 24 months. 6 Roughly one third of all patients reached this primary endpoint and two thirds achieved normalized ALP values with bezafibrate. The main side effects in the bezafibrate group were a 5% increase in creatinine values due to a pharmacological effect, not caused by reduced kidney function, and myalgia in 20% of patients (10% of patients had myalgia in the placebo group). 6 Although bezafibrate has not yet been approved for the treatment of PBC, it is an appealing offlabel treatment option in non-responders to UDCA treatment, with a favourable side-effect profile and low treatment costs. A recent small trial from Leuven, Belgium, evaluated the additive effects of bezafibrate in patients who did not respond adequately to dual therapy with OCA and UDCA. Triple therapy (bezafibrate, OCA, UDCA) lowered cholestatic parameters and improved pruritus in most patients. 7 There are ongoing trials to clarify the mechanism of action of bezafibrate, which is an agonist for the peroxisome proliferator-activated receptor (PPAR). While this agent works as an agonist on different PPARs, selective PPAR-δ-(Seladelpar; CYMABAY) or PPAR-α/δ-agonists (Elafibranor; GENFIT) are also under evaluation in phase III trials for PBC to improve the efficacy and reduce side effects in patients who do not respond to UDCA therapy. Seladelpar demonstrated anti-cholestatic properties and improved pruritus in both patients with and without cirrhosis (Child A) 8 and elafibranor has also been shown to have anti-cholestatic effects associated with a reduction in pruritus. 9 Another trial in PBC is evaluating the use of the fibroblast growth factor 19 (FGF19) agonist NGM282 (NGMBio). FGF19 is mainly located in the ileum and serves to regulate bile acid metabolism.
Production of FGF19 is induced by bile acid-dependent activation of the FXR receptor. FGF19 translocates via the portal system to the liver and suppresses expression of the rate-limiting enzyme for de novo bile acid synthesis in the liver (CYP7A1). In a small 28-day trial, NGM282 improved ALP by at least 15% from baseline. 10 In summary, PBC, as a role model for autoimmune diseases in general and liver diseases in particular, is no longer a major indication

| PRIMARY SCLEROS ING CHOL ANG ITIS
Primary sclerosing cholangitis is a rare liver disease that significantly affects quality of life, morbidity and mortality. The aetiology of PSC is unknown and its pathogenesis is poorly understood. The only existing life-saving treatment option is liver transplantation. Nevertheless, PSC reoccurs in 20%-37% of patients after liver transplantation. 11 No drugs have been shown to be effective in preventing death, liver transplantation or cholangiocarcinoma (CCA). Although UDCA improves ALP, it does not improve survival. The role of ALP as a surrogate marker for treatment response is a subject of debate. The prognosis is poor in patients with PSC; up to 40% require liver transplantation and up to 31% develop CCA. Inflammatory bowel disease (IBD) is present in most cases. 12,13 The diagnosis of PSC is mainly based on imag- It is also urgent to improve the risk stratification of unfavourable outcomes in patients with PSC. Recent studies have shown that elevated Immunoglobulin G (IgG) at diagnosis is an independent predictor of transplant-free survival. 15 The Amsterdam-Oxford model has been proposed to stratify the risk of liver transplantation or death in PSC patients using seven available variables (albumin, platelet count, age, PSC subtype, aspartate-aminotransferase, ALP and bilirubin). 16 New biomarkers are needed to diagnose PSC, predict disease outcome and serve as monitoring parameters. Bile proteomic profiles might help identify PSC patients at risk of developing CCA, thus facilitating early diagnosis of CCA and improving the prognosis by reinforcing screening for malignancy. 17 Because UDCA has not been found to improve the outcome of liver disease in patients with PSC, new therapies are urgently needed. Recent evidence has supported a link between the liver and the gut in the pathogenesis of PSC. The potential impact of IBD therapies on PSC were evaluated based on the strong association between PSC and IBD. A retrospective analysis by Tse et al found initial improvement in ALP in patients with PSC and concomitant IBD who were treated with the anti-tumour-necrosis factor alpha (TNF) antibody, adalimumab. 18 However, these findings were not confirmed when PSC patients were treated with another anti-TNF antibody, infliximab. No improvement in liver biochemistry was seen in patients with IBD and concomitant PSC receiving vedolizumab, an α4β7 integrin inhibitor. 18 A recent trial investigated the feasibility and safety of faecal microbiota transplantation (FMT) in patients with PSC. Ten patients with PSC and concomitant IBD and an ALP >1.5 × ULN were included and received a single FMT via colonoscopy. A more than 50% decline in ALP levels was observed in 30% of patients and no adverse events occurred. 19 Thus, further evaluation of FMT is needed in patients with PSC and IBD. The non-steroidal FXR-agonist, cilofexor has been shown to improve cholestasis and liver injury in a small trial with 52 patients who were randomized to receive a placebo and two groups with different doses of cilofexor. Treatment with the maximum dose of cilofexor over a period of 12 weeks led to a 21% decrease in ALP and to significant improvement in aminotransferase levels. 20 Nor-UDCA, a side-chain modified derivate of UDCA, had dose-dependent beneficial effects on ALP values in a phase II trial. 21

| AUTOIMMUNE HEPATITIS
Autoimmune hepatitis is characterized by elevated aminotransferases, hypergammaglobulinaemia, the presence of characteristic auto-antibodies (e.g. anti-nuclear antibodies [ANA] and smooth muscle antibodies [SMA] in AIH type 1, liver-kidney-microsome antibodies [LKM] in AIH type 2) and histopathological features such as plasma cell enriched infiltrates, emperipolesis and interface hepatitis. Neither histopathology, auto-antibodies or hypergammaglobulinaemia alone can diagnose AIH. All of these features and the exclusion of other liver diseases are necessary to make the diagnosis of this disease, which can be confirmed by a response to corticosteroids. New diagnostic markers are needed to diagnose AIH as well as to monitor treatment response. Existing auto-antibodies either lack sensitivity or specificity or both. Although the sensitivity and specificity of ANA and SMA are acceptable, they are not specific for AIH. Anti-solubleliver-antigen antibodies and LKM-antibodies are highly specific but their sensitivity is limited. 23 Although anti-asialoglycoprotein receptor antibodies have the best overall performance, they are not widely available. 24 We recently showed that anti-huntingtin-interacting protein 1-related protein (anti-HIP1R protein) antibodies, measured using enzyme-linked immunosorbent assay, were elevated in AIH patients with a higher specificity than and equal sensitivity to ANA and SMA. 25 Normalization of IgG and aminotransferases as well as a lack of inflammatory activity in liver histology are the main treatment goals, called complete remission. A subgroup of patients fails to achieve remission with corticosteroids and/or azathioprine, which are the only approved agents for the treatment of AIH. Large randomized, controlled trials that evaluate the efficacy and safety of other immunosuppressive drugs except for budesonide, are lacking. 26 27 infectious complications developed in up to 70% of these patients. However, several case reports indicate that TNF-antibodies may even cause AIH. 28 Rituximab, a B-cell depleting agent, has been shown to induce remission in hard-to-treat AIH patients with lower complication rates. 29 Another B-cell depleting monoclonal antibody targeting the B-cell activation factor receptor is under investigation in a phase II/III trial in AIH patients who do not respond to SOC. 30 Furthermore, there is a reduced balance of regulatory and effector T cells in patients with AIH who do not achieve remission. Interleukin-2 (IL-2) is a key cytokine for T-cell tolerance and IL-2 therapy has shown to restore immune tolerance by restoring impaired regulatory T-cell function. A case report of low dose IL-2 therapy with 1 million units five times a month for six months showed an increase in regulatory T cells and induced remission in one of two treated AIH patients. 31 Data on the transfer of regulatory T cells in humans for this indication will soon be available. Both IL-2 therapy and the transfer of regulatory T cells are promising strategies, which require further investigation in larger trials. Because AIH is a rare disease and most patients respond to first line therapy, large prospective multicentre studies are difficult. Nevertheless, a multicentre international trial is urgently needed to perform robust studies for salvage therapies in patients with AIH and an incomplete response or intolerance to SOC. The European Reference Network (ERN) for rare liver diseases could facilitate the implementation of these studies in the future.

| SUMMARY
Overall, PBC is probably the best understood and well managed au-

ACK N OWLED G EM ENTS
None.

Bastian Engel declares no conflict of interest related to this work. Richard
Taubert is inventor of the patent application for the use of anti-HIP1R for the diagnosis of AIH. Elmar Jaeckel is inventor of the patent application for the use of anti-HIP1R for the diagnosis of AIH. Michael P. Manns is or was principal investigator for Falk Pharma, Gilead, Intercept and Novartis. He received grants from Falk Pharma, Gilead and Intercept. Michael Manns is or was a consultant for Falk Pharma, Gilead, Intercept, Roche and Novartis and he lectured for Falk Pharma and Gilead. Travel