REVIEWS
Genetics of primary sclerosing cholangitis and pathophysiological implications Xiaojun Jiang1–4 and Tom H. Karlsen1–4
Abstract | Primary sclerosing cholangitis (PSC) is a chronic disease leading to fibrotic scarring of the intrahepatic and extrahepatic bile ducts, causing considerable morbidity and mortality via the development of cholestatic liver cirrhosis, concurrent IBD and a high risk of bile duct cancer. Expectations have been high that genetic studies would determine key factors in PSC pathogenesis to support the development of effective medical therapies. Through the application of genome-wide association studies, a large number of disease susceptibility genes have been identified. The overall genetic architecture of PSC shares features with both autoimmune diseases and IBD. Strong human leukocyte antigen gene associations, along with several susceptibility genes that are critically involved in T-cell function, support the involvement of adaptive immune responses in disease pathogenesis, and position PSC as an autoimmune disease. In this Review, we survey the developments that have led to these gene discoveries. We also elaborate relevant interpretations of individual gene findings in the context of established disease models in PSC, and propose relevant translational research efforts to pursue novel insights.
Correspondence to T.H.K. Norwegian PSC Research Center, Department of Transplantation Medicine, Division of Cancer Medicine, Surgery and Transplantation, Oslo University Hospital Rikshospitalet, Postbox 4950 Nydalen, 0424 Oslo, Norway.
[email protected] doi:10.1038/nrgastro.2016.154 Published online 15 Mar 2017; corrected online 30 Mar 2017
Primary sclerosing cholangitis (PSC) is a chronic, fibrotic bile duct disease with an estimated prevalence in Western countries of ~1 per 10,000 (REFS 1,2). In Asia, the preva‑ lence seems to be ~10‑fold lower3. Ulcerative colitis is the most frequently classified bowel affection in patients with PSC, although some patients are also categorized by clinical features as having Crohn’s disease colitis4–6. The biliary affection in PSC is chronic and represented by mild inflammation and progressive concentric fibrosis (‘onion skin’ fibrosis) around the bile duct (FIGS 1,2). The relationship between bowel and bile duct manifestations of PSC is unclear, with bile duct disease even preced‑ ing bowel affection in some patients7. Bowel disease in PSC is managed according to standard regimens for IBD8, with heightened awareness of the risk of colorectal malignancy9. For biliary affection, no medical therapy has so far been shown to influence disease progres‑ sion10. Most patients show a progressive disease course, resulting in a median time until liver transplantation is required of 10–20 years (depending on cohort character istics, secondary versus primary referral centres)11,12. In Scandinavia, PSC is the leading cause of liver trans‑ plantation13, and in all Western countries represents the leading indication for liver transplantation within the spectrum of cholestatic liver disease14.
From a genetic terminology perspective, the pheno typic presentation of PSC shows important overlap with other diseases. The degree of comorbidity with IBD is considerable, but shows important geographical vari ation. In Northern Europe and the USA, up to 60–80% of patients with PSC also have a clinical diagnosis of IBD15. In Southern Europe and Asia, this fraction is lower (30–50%)16–19. A large number of patients in Northern Europe and the USA (~25%) also have autoimmune comorbidities outside of the gut–liver axis, for e xample autoimmune thyroid disease, type 1 diabetes melli‑ tus (T1DM) and rheumatoid arthritis20. Furthermore, the high lifetime risk of bile duct and gallbladder cancer (up to 15%) poses additional challenges for clinical management21, and currently comprises a major cause of death in patients with PSC22. The phenotypic heterogeneity has led to specula‑ tions that PSC might comprise a ‘mixed bag’ of hitherto undefined conditions, hence explaining the difficulty in delineating the aetiology and pathogenesis of the dis‑ ease. Geographical variability of the clinical comorbid ities supports such notions, yet for the major subgroup of patients who have PSC in the context of IBD, a common pathophysiological basis probably exists. Even in Japan23, where IBD frequency for patients with PSC is reported
NATURE REVIEWS | GASTROENTEROLOGY & HEPATOLOGY
VOLUME 14 | MAY 2017 | 279
. d e v r e s e r s t h g i r l l A . e r u t a N r e g n i r p S f o t r a p , d e t i m i L s r e h s i l b u P n a l l i m c a M 7 1 0 2 ©
REVIEWS Key points • Primary sclerosing cholangitis (PSC) is a complex genetic disease of the bile ducts and the bowel, in which multiple genetic and environmental factors interact in causing inflammation and fibrosis • 22 susceptibility loci for PSC have been established at a genome-wide significance level (P ≤5 × 10−8), with the HLA complex representing the strongest finding by several orders of magnitude • The overall genetic architecture of PSC resembles that of prototypical autoimmune diseases, with some genetic overlap also being evident for IBD susceptibility genes • Genetic findings in PSC so far explain 50% of the unexplained fraction • Individual gene findings should be interpreted with an emphasis on tissue-specific and PSC-specific functions, and published studies performed before the knowledge of a PSC association might not be relevant • The pool of susceptibility genes serves a major resource for translational studies aimed at deciphering pathophysiology in PSC, and could guide developments in the direction of effective treatments
to be ~30%, patients with onset of the disease at a young age (20–40 years) exhibit similar IBD frequencies as in Europe (almost 60%) and, therefore, probably repre‑ sent a PSC subset similar to that observed in Western countries24. The proportion of total patients with IBD represented by this IBD-associated sclerosing cholangitis entity has been established at 8.1%25, but less than half of these cases are clinically apparent. Compared with ulcerative colitis and Crohn’s disease, PSC is associated with particular clinical features of IBD (pancolitis, rightsided colitis, rectal sparing, ileitis and increased risk of colorectal cancer)26. In this Review, we aim to summarize the develop‑ ments in PSC genetics over the years. An emphasis will be placed on the interpretation of findings and poten‑ tial pathophysiological implications. Furthermore, we will elaborate on the relationship between the genetics of PSC and those of other inflammatory conditions, demonstrating how the findings from genetic studies have reinforced PSC as a separate disease entity, over and above that of a complication occurring in subsets of patients with ulcerative colitis or Crohn’s disease.
Liver disease genetics and GWAS During the 1990s, robust disease gene discoveries in liver disease genetics were limited to Mendelian traits, starting with the hyperbilirubinaemias and Wilson disease27–32, with a subsequent focus on hereditary cholestasis and haemochromatosis33–43. This era and the interpretation of the gene findings in these studies has greatly influenced the understanding of susceptibility genes in non-Mendelian (that is, complex) diseases such as PSC. This point is important to be aware of, because the genetic findings to be reviewed in this article rep‑ resent different mechanisms of causality. In Mendelian diseases, there is an apparent 1:1 relationship between genetic aberrations and disease traits (that is, the genetic variants cause disease), although penetrance is rarely 100% due to the variable presence of environmental and genetic factors that interact with the disease mutation and modify risk. However, in complex disease genet‑ ics, contextual factors (such as the environment and
gene–gene interactions) play a considerably greater part, making it inappropriate to assume susceptibility genes are causal. Thus, to claim that each of the PSC-associated genetic variants cause disease is inappropriate44. The distinction between Mendelian genetics and complex disease genetics is underscored by the fact that the over‑ all contribution of genetics to complex traits like PSC is limited45. In complex diseases, genetic findings, even by the most comprehensive mathematical extrapolations, are unlikely to represent much more than 30–40% of overall disease liability46,47. The first successful genome-wide association study (GWAS) was published in 2005 (REF. 48), and the follow ing decade saw many applications of the study design, leading to the identification of >1,000 risk loci for a variety of complex human traits. A GWAS is, in simple terms, a statistical analysis in which the genetic vari‑ ants spread throughout the genome are assessed for their relationship with a particular feature (phenotype), such as the presence of PSC. For disease phenotypes, risk loci (susceptibility loci) are defined as chromosomal regions (sometimes affecting a single susceptibility gene) for which there is a statistically significant difference in the occurrence of particular variants observed in patients compared with healthy individuals as controls. Notably, it might not be possible to attribute a specific risk gene to an identified risk locus owing to a genetic phenomenon called linkage disequilibrium. Linkage disequilibrium is a statistical measure of correlations between neighbour‑ ing genetic variants in a given population arising because chromosomal segments are inherited in large blocks, despite some recombination occurring during meiosis49. For some risk loci, for example the macrophage stimu‑ lating 1 (MST1) locus on 3p21 and the human leukocyte antigen (HLA) complex on chromosome 6p21, linkage disequilibrium extends over a particularly large number of theoretically relevant genes50,51. Owing to the high number of genetic variants tested when performing a GWAS (typically ~1,000,000 for most modern genotyping arrays), the statistical significance threshold is stringently set by convention at P ≤5 × 10−8 (so-called genome-wide significance) to avoid falsepositive findings (type 1 statistical errors)52. Inherent to the association analysis study design, for variants at risk loci to be detected they must have a relatively high Author addresses Norwegian PSC Research Center, Department of Transplantation Medicine, Division of Cancer Medicine, Surgery and Transplantation, Oslo University Hospital Rikshospitalet, Postbox 4950 Nydalen, 0424 Oslo, Norway. 2 Institute of Clinical Medicine, Faculty of Medicine, University of Oslo, Postbox 1171 Blindern, 0318 Oslo, Norway. 3 Research Institute of Internal Medicine, Division of Cancer Medicine, Surgery and Transplantation, Oslo University Hospital Rikshospitalet, Postbox 4950 Nydalen, 0424 Oslo, Norway. 4 K.G. Jebsen Inflammation Research Centre, Institute of Clinical Medicine, Faculty of Medicine, University of Oslo, Postbox 1171 Blindern, 0318 Oslo, Norway. 1
280 | MAY 2017 | VOLUME 14
www.nature.com/nrgastro . d e v r e s e r s t h g i r l l A . e r u t a N r e g n i r p S f o t r a p , d e t i m i L s r e h s i l b u P n a l l i m c a M 7 1 0 2 ©
REVIEWS Normal
Healthy bile duct
Bile
Disease progression
Preclinical
Cholangiocyte
Biliary bicarbonate umbrella
Hepatic stellate cell
Macrophage Lymphocyte Neutrophil Fibroblast
Portal myofibroblast
Cilia
Stricture
Pathogenic processes Preclinical phenotype • Oxidative and nonspecific stress • Microbial dysbiosis • Antigenic T-cell activation • Viral infection • Cholangiocyte activation
Cholangiocyte proliferation
Peribiliary gland Biliary stem cell Early lesion • Inflammation • Cholangiocyte activation • Peribillary gland enhancement
Onion skinning fibrosis
Typical PSC • Inducible myofibroblasts and hepatic stellate cell activation • Fibrosis (‘onion skinning’) • Dysplasia and cholangiocarcinoma
Nature Reviews | Gastroenterology Hepatology Figure 1 | The principle aspects of bile duct lesions in PSC. Primary sclerosing cholangitis (PSC) is a slowly &progressive disease that conceptually can be graded according to a preclinical state, an early lesion and the typical lesion with ‘onion skinning’ fibrosis (appearing as concentric layers of rings of fibrosis on histology). Importantly, multiple different stages of biliary affections typically occur side by side in the affected liver (so‑called ‘patchy’ affection).
frequency (that is, they are ‘common’, typically with a frequency >5% in the general population). Common variants generally have a low effect on biological traits, with a typical odds ratio below 1.3 or 1.4 (REF. 53). The latter fact implies that large numbers (prefer ably t housands) of cases and controls are required to achieve sufficient statistical power. For the study design to be useful, much effort has gone into establishing researcher networks to recruit patients for DNA collec‑ tion. This collaborative, international working environ‑ ment should be considered a beneficial ‘side-effect’ of GWAS54. For rare diseases such as PSC, the statistical stringency and the low effect size of implied variants inevitably leads to false-negatives (type II statistical errors) owing to small study populations, and this prop‑ erty has to be kept in mind as a limitation of the studies reviewed here. Susceptibility genes identified so far explain
