REVIEW URRENT C OPINION
Update on primary sclerosing cholangitis genetics Eva K.K. Henriksen a,b, Espen Melum a,b, and Tom H. Karlsen a,b,c
Purpose of review The pathogenesis of primary sclerosing cholangitis (PSC) involves heritable factors. This review summarizes the recent genetic studies and discusses the implications of identified risk loci. Recent findings A total of 16 PSC susceptibility loci have been identified in genome-wide association studies and related study designs. At least 33 additional loci are involved in what is increasingly acknowledged to represent a general pool of genetic risk loci for immune-mediated diseases. One important group of genes is part of well characterized immune pathways (e.g. interleukin 2 signaling), whereas for other loci the relationship to PSC pathophysiology is less evident. Importantly, the loci collectively account for only 7.3% of overall PSC liability, thus pointing to a large contribution from environmental factors to PSC development. The individual PSC risk genes cannot be interpreted within a simple cause–effect model used for monogenic traits, but need to be explored for their individual biological correlates, preferably in a disease context. To some extent, as exemplified for the human leukocyte antigen and FUT2 associations, genetic findings may guide the discovery of interacting and co-occuring environmental susceptibility factors. Summary Multiple PSC susceptibility loci are now available for exploration in experimental model systems and patient-centered research. Keywords genome-wide association study, primary sclerosing cholangitis
INTRODUCTION In genetic terms, primary sclerosing cholangitis (PSC) is a complex phenotype, that is a phenotype caused by interplay of genetics and environment. The heritable contribution to PSC is comparable to most other autoimmune and inflammatory diseases, with PSC occurring in first-degree relatives of patients slightly more than 10 times as frequently as in the general population [1]. Recent genome-wide association studies (GWASs) (Table 1) [2–4,5 –7 ] provide an increasing level of clarity on the role of the specific genetic factors involved in this heritability. At the time of writing, a total of 16 PSC risk loci (Table 2) [8–48] have been reported at a significance threshold robust to strict statistical correction of multiple testing and risk of type 1 errors (i.e. genome-wide significance, P < 5 108). Several additional disease loci (at least 33) likely exist (Table 3) [7 ]. Investigators are already perusing this information in designing experiments for elucidation of PSC pathogenesis. But what information pertains? In the present review, we will survey the current knowledge on genetics of PSC with a particular emphasis on this question, and discuss how the answer has &&
implications for post-GWAS translational and genetic research.
THE UPDATED GENE LIST AND PLEIOTROPY In principle, two broad categories of risk loci have been shown to associate with PSC so far. First, there is a category of risk loci involving genes encoding molecules serving key components of adaptive and
&&
&&
www.co-gastroenterology.com
a Division of Cancer Medicine, Surgery and Transplantation, Department of Transplantation Medicine, Norwegian PSC Research Center, Oslo University Hospital Rikshospitalet, bDivision of Cancer Medicine, Surgery and Transplantation, K.G. Jebsen Inflammation Research Centre, Research Institute of Internal Medicine, Oslo University Hospital Rikshospitalet, Oslo and cDepartment of Clinical Medicine, University of Bergen, Bergen, Norway
Correspondence to Professor Tom H. Karlsen, MD, PhD, Division of Cancer Medicine, Surgery and Transplantation, Department of Transplantation Medicine, Norwegian PSC Research Center, Oslo University Hospital Rikshospitalet, Postboks 4950 Nydalen, N-0424 Oslo, Norway. Tel: +47 2307 2469; fax: +47 2307 3928; e-mail: t.h.karlsen@medisin. uio.no Curr Opin Gastroenterol 2014, 30:310–319 DOI:10.1097/MOG.0000000000000052 Volume 30 Number 3 May 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Update on primary sclerosing cholangitis genetics Henriksen et al.
KEY POINTS A total of 16 risk loci have been identified in PSC at genome-wide significance level. The 16 risk loci collectively account for 7.3% of the overall PSC liability. Environmental factors interact with several of the identified risk loci and make up the majority of the overall PSC liability.
individual loci. It is important to acknowledge that the concepts coming from the genetics of monogenic cholestatic traits (e.g. progressive familial intrahepatic cholestasis and cystic fibrosis cholangiopathy) are not valid when assessing the PSC genetics. The risk variants associating with PSC do not occur in all patients, and, indeed, occur in the healthy population even in the homozygous form. Two striking features are apparent when considering the loci collectively. First, as reviewed elsewhere [49 ], all loci show associations in other prototypical autoimmune diseases such as type 1 diabetes and multiple sclerosis. Secondly, half of the loci show significantly stronger associations in PSC than in inflammatory bowel disease (IBD) [7 ]. Deriving from the profound overlap in genetic associations between immune-mediated diseases, in genetic terminology called pleiotropy, the initiating factors in PSC pathogenesis are likely to share features with those known to involve in these genetically related diseases. The pleiotropic findings have implication to the ongoing disputes as to whether a toxic bile duct insult or immune-mediated bile duct destruction serves as the initiating event in PSC. Although from genetic data, it seems likely that immune perturbations occur prior to bile acid toxicity, the two schools are not mutually exclusive. Throughout disease course in PSC, cholestasis and dysregulation of bile acid metabolism and transport increase, and there is also cross-talk &
Experimental research on identified loci needs to be incorporated with the existing models of sclerosing cholangitis and patient data.
&&
innate immunity. These genes include IL2/IL21, IL2RA, BACH2, TNFRSF14, CD28 and CD226. The second category of loci involves genes on which knowledge on function exists, but for which the relationship to immune function and PSC pathogenesis is less clear. These genes include BCL2L11, MST1, HDAC7, SH2B3, TCF4, SIK2, GPR35, PRKD2 and PSMG1. It is useful to provide separate considerations for the chromosome 6p21 region, known to harbor the major histocompatibility complex that encodes the human leukocyte antigen (HLA) class I and class II molecules. The most fruitful interpretations of the outcome of genetics in PSC so far are not made on the basis of
Table 1. List of recent genetic studies reporting on genome-wide significant risk loci (P < 5 108) in primary sclerosing cholangitis N patients
Study (year)
N controls
Participating countries
GWS loci
Karlsen et al. [2] (2010)
285
298
Norway
MHC
Melum et al. [3] (2011)
1740
5136
Norway, Sweden, Germany, the Netherlands, Belgium and the United States
MHC, BCL2L11 and MST1
992
5162
United Kingdom (primary study) along with Norway, Sweden, Germany, the Netherlands, Belgium and the United States (for the meta-analysis)
MST1 and IL2RA
1936
6470
Norway, Sweden, Germany, the Netherlands, Belgium and the United States
TNFRSF14/MMEL1
Srivastava et al. [4] (2012)
Folseraas et al. [5 ] (2012) &&
Ellinghaus et al. [6 ] (2013)
1401
5530
Germany, Norway and Sweden
GPR35 and TCF4
Liu et al. [7 ] (2013)
3789
25079
Pan European, the United States and Canada
MHC, MMEL1-TNFRSF14, CD28, MST1, IL2/IL21, BACH2, IL2RA, SIK2, HDAC7, SH2B3/ATXN2, CD226, PRKD2/STRN4 and PSMG1
&&
&&
GWS, genome-wide significant; MHC; major histocompatibility complex. For gene name abbreviations, see http://www.ncbi.nlm.nih.gov/gene/.
0267-1379 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
www.co-gastroenterology.com
311
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Biliary tract Table 2. List of the 15 primary sclerosing cholangitis susceptibility loci identified at genome-wide significance (P < 5 108) outside of the major histocompatibility complex (chromosome 6p21, the 16th locus)
Chromosome
Main candidate gene(s)
Human monogenic traits
1p36
TNFRSF14
–
Diminished inflammatory response to DSS [8]; more susceptible to EAE [9]; decreased osteoclastogenesis via RANKL [10]; reduced severity of HSV-1 keratitis [11]; impaired renal Epo response to erythropoietic stress [12]; reduced body weight gain on high-fat diet and reduced adipose tissue macrophage and T-cell recruitment [13]; attenuated hepatitis after ConA injection [14], but increased morbidity in response to ConA in old mice [9]
–
MMEL1
–
Reduced fertility (male) [15]; elevations of total bamyloid in the hippocampus and diencephalon/ brainstem [16]
Racecadotril
2q13
BCL2L11
–
Accumulation of lymphoid and myleoid cells, and reduction of platelet number [17]; on the C57BL/6 129SV genetic background: most exhibit accumulation of plasma cells and development of progressive systemic autoimmune disease, including autoimmune kidney disease and lupus erythematosus [17]; protection from hepatitis after LCMV infection [18]; protection from anti-Fas induced liver damage [19]
Cyclophosphamide, glucocorticoids and HDAC inhibitors such as panobinostat (indirect)
2q33
CD28
–
Reduced basal immunoglobulin levels [20]; diminished Ig class switching after infection with vesicular stomatitis virus [20]
Abatacept and belatacept
2q37
GPR35
–
Increased systemic blood pressure [21]
Lodoxamide and aspirin
3p21
MST1
–
Lipid-filled cytoplasmic vacuoles in hepatocytes [22]
Crizotinib (indirect)
4q27
IL2
–
On BALB/c genetic background: generalized autoimmune disease, preferentially hemolytic anemia [23]; on 129/Ola x C57BL/6 genetic background: rapid onset of disease symptoms, with 50% mortality by 9 weeks; survivors develop ulcerative colitis-like disease with an age-related reduction in the number of B cells [24–26]; germ-free mice develop milder gastrointestinal inflammation [27]; inhibited LCMV-induced T cell expansion [28]
Tacrolimus, cyclosporine, calcitriol and corticosteroids (indirect)
IL21
–
Enhanced IgE isotype switch, expansion of IgE positive cells and IgE production [29]
NNC0114-0005
6q15
BACH2
–
Impaired differentiation of B cells and reduced B cell numbers [30]; increased production of IgM, but diminished concentrations of IgG subclasses and IgA [30]; deficient IgG responses associated with severe reduction in somatic hypermutation and class switch recombination [30]; progressive wasting disease and reduced survival [31,32]; PAP-like accumulation of surfactant proteins in the lungs and modest intestinal inflammation [31,32], accompanied by increased levels of autoantibodies [32]
Heme arginate (indirect)
10p15
IL2RA
Deficiency of interleukin 2 receptor a (OMIM606367)
Massive enlargement of peripheral lymphoid organs associated with expansion of B cells and T cells [33]; older mice develop autoimmune disorders, including inflammatory bowel disease and hemolytic anemia [33]
Aldesleukin, denileukin diftitox, daclizumab and basiliximab
312
www.co-gastroenterology.com
Main biological alterations observed in murine knockout models
Examples of relevant pharmacology
Volume 30 Number 3 May 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Update on primary sclerosing cholangitis genetics Henriksen et al. Table 2 (Continued)
Chromosome
Main candidate gene(s)
Human monogenic traits
Main biological alterations observed in murine knockout models
11q23
SIK2
–
Protection from ischemic brain injury [34]
Glucagon (indirect)
12q13
HDAC7
–
Embryonic lethality due to vascular dilation and rupture caused by endothelial cell adhesion defects [35]; conditional deletion in the osteoclast lineage promotes osteoclastogenesis and bone resorption [36]
Valproic acid
12q24
SH2B3
Hematological malignanciesa
Changes in the hematopoietic stem cell compartment, including abnormal lymphoid and myleoid homeostasis and splenomegaly [37]; develop B cell precursor acute lymphoma following irradiation [38]
–
ATXN2
Spinocerebellar ataxia-2 (OMIM183090)
Reduced fertility, mild defect in motor learning and abnormal fear-related behaviors [39–41]; enlarged abdominal fat pad and seminal vesicles [40]; adult-onset obesity, reduced insulin receptor expression in cerebellum and liver, and increased insulin levels in pancreas and blood serum [40,41]; dyslipidemia and liver steatosis with accompanying cholestasis [40,41]
–
18q21
TCF4
Pitt–Hopkins syndrome (OMIM610954)
Die within 1 week after birth [42,43]; reduced capacity to generate pro-B cells [42]; disrupted pontine nucleus development [43]; specific reduction of plasmacytoid dendritic cells after activation of conditional allele [44]
–
18q22
CD226
–
Increased tumor formation and mortality following transplantation of CD155-expressing cells [45]; enhanced development of fibrosarcoma and papilloma in response to chemical carcinogens [45]; delayed clearance of LCMV [46]
Doxorubicin and melphalan (indirect)
19q13
PRKD2
–
Gene trap deletion results in no apparent phenotype [47]
–
STRN4
–
NA
Tamoxifen and trifluoperazine (indirect)
PSMG1
–
Early embryonic lethality [48]; conditional deletion in the liver results in premature hepatocyte senescence [48]; conditional deletion in the brain causes growth retardation and abnormal brain development [48]
–
21q22
Examples of relevant pharmacology
Along with each locus, key biological alterations associated with loss of gene function in humans and mice are given along with known examples of pharmacological interactions with affected pathways. Candidate genes and variants are not unambiguously identified for any of the loci, but are given for simplicity and to allow for interpretation of findings. a Somatic mutations. ConA, Concanavalin A; DSS, dextran sodium sulphate; EAE, experimental autoimmune encephalomyelitis; HSV, herpes simplex virus; LCMV, lymphocytic choriomeningitis virus; NA, not applicable; PAP, pulmonary alveolar proteinosis. The OMIM number refers to the access number in the Online Mendelian Inheritance in Man database, see http://www.ncbi.nlm.nih.gov/omim. For gene name abbreviations, see http://www.ncbi.nlm.nih.gov/gene/.
between immune regulation and these aspects [50,51]. The limited overlap between PSC and IBD in the most recent genetic analysis was surprising [7 ]. That half of the loci do not show robust associations in IBD indicate that PSC is more distinct from IBD than ulcerative colitis is from Crohn’s disease (ulcerative colitis and Crohn’s disease share at least 70% of the &&
&
loci [52 ]). At the level of statistical power exhibited by the PSC Immunochip study, it would have been expected that more of the 163 IBD loci also associate with PSC given the frequency of IBD in the study population of 72%. When GWASs on IBD were reporting on patient populations sized approximately at n ¼ 1500–3000, the number of loci detected was between 13 and 30 [53–56]. If IBD in PSC was
0267-1379 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
www.co-gastroenterology.com
313
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Biliary tract Table 3. Pleotropic risk loci identified in the primary sclerosing cholangitis Immunochip study SNP
Main candidate gene(s)
Genes in region
rs4474277
SDHC
rs12479056
PUS10, REL
14
CD, CeD, Psor, RA, Sarc, T1D,UC
rs11676348
TGR5, CXCR1/2
10
CD, CeD, Psor, RA, Sarc, T1D,UC
3
rs7556897
Locus supported by phenotype RA
0
CD, Psor, RA, Sarc, T1D,UC
rs9819066
FOXP1
2
CD, CeD, Psor, RA, Sarc, T1D,UC
rs4482697
IQCB1
7
CeD
rs7636495
LRRC33
1
CeD
rs898518
LEF1
5
CeD
rs9378805
IRF4
1
CeD, RA, T1D
rs12210050
EXOC2
1
CeD, Psor, RA, Sarc, T1D, UC
rs535780
PRDM1
1
CD, Psor, RA, Sarc, T1D, UC
rs10956390
PVT1
3
CD, RA, UC
rs13255292
MIR1207
2
RA, T1D
rs2977035
MIR1208
1
CD, CeD, Psor, RA, Sarc, UC
rs6477901
ROD1
5
CD, CeD, Psor, RA, Sarc, T1D, UC
rs7923837
HHEX
3
CD, CeD, Psor, RA, Sarc, T1D, UC
rs10883371
NKX2-3
1
CD, CeD, Psor, RA, T1D, UC
rs11246286
CD151
18
CD, CeD, Psor, RA, Sarc, T1D, UC
rs694739
PRDX5
21
CD, CeD, Psor, RA, Sarc, T1D, UC
rs633683
CXCR5
9
CD, CeD, Psor, RA, Sarc, T1D, UC
rs12369214
RIC8B
7
CD, CeD, Psor, RA, Sarc, T1D, UC
rs7324586
MIR548F1
1
CD, Psor, UC
rs4983425
C14orf80
7
Psor, RA, UC
rs415595
SOCS1
6
CD, CeD, Psor, RA, Sarc, T1D, UC
rs7404095
PRKCB
2
UC
rs12149608
ZFP90
4
Psor, RA, Sarc, UC
rs4795397
ZPBP2
24
rs2847297
PTPN2
1
CD, CeD, Psor, RA, T1D, UC
rs17694108
SLC7A10
1
Psor, RA, Sarc, UC
rs601338
FUT2
rs715147
NFATC2
2
RA
rs11203203
UBASH3A
3
CD, CeD, Psor, RA, Sarc, T1D, UC
rs2838519
ICOSLG
5
CD, CeD, Psor, RA, Sarc, T1D, UC
15
RA
CD, Psor, Sarc
The table lists 33 loci achieving a false discovery rate