CCR5-A32 Mutation Is Strongly Associated with Primary Sclerosing Cholangitis

FULL PAPER CCR5-D32 mutation is strongly associated with primary sclerosing cholangitis

R Eri1, JR Jonsson2, N Pandeya3, DM Purdie3, AD Clouston2, N Martin4, D Duffy4, EE Powell2,5, J Fawcett6, THJ Florin1,7 and GL Radford-Smith1,8 1Brisbane IBD Research Group, Clinical Research Centre, Royal Brisbane Hospital Research Foundation, Brisbane, Australia; 2University of Queensland School of Medicine, Southern Division, Princess Alexandra Hospital, Brisbane, Australia; 3Queensland Institute of Medical Research, Population and Clinical Sciences Division, Brisbane, Australia; 4Queensland Institute of Medical Research, Genetic Epidemiology Unit, Brisbane, Australia; 5Department of Gastroenterology and Hepatology, Princess Alexandra Hospital, Brisbane, Australia; 6Queensland Liver Transplant Service, Brisbane, Brisbane, Australia; 7University of Queensland Department of Medicine, Mater Health Services Adult Hospital, Brisbane, Australia; 8Department of Gastroenterology, Ned Hanlon Building, Royal Brisbane and Women’s Hospital, Brisbane, Australia

CCR5 plays a key role in the distribution of CD45RO þ T cells and contributes to generation of a T helper 1 immune response. CCR5-D32 is a 32-bp deletion associated with significant reduction in cell surface expression of the receptor. We investigated the role of CCR5-D32 on susceptibility to ulcerative colitis (UC), Crohn’s disease (CD) and primary sclerosing cholangitis (PSC). Genotype and allelic association analyses were performed in 162 patients with UC, 131 with CD, 71 with PSC and 419 matched controls. There was a significant difference in CCR5 genotype (OR 2.27, P ¼ 0.003) between patients with sclerosing cholangitis and controls. Similarly, CCR5-D32 allele frequency was significantly higher in sclerosing cholangitis (17.6%) compared to controls (9.9%, OR 2.47, P ¼ 0.007) and inflammatory bowel disease patients without sclerosing cholangitis (11.3%, OR 1.9, P ¼ 0.027). There were no significant differences in CCR5 genotype or allele frequency between those with either UC or CD and controls. Genotypes with the CCR5-D32 variant were increased in patients with severe liver disease defined by portal hypertension and/or transplantation (45%) compared to those with mild liver disease (21%, OR 3.17, P ¼ 0.03). The CCR5-D32 mutation may influence disease susceptibility and severity in patients with PSC. Genes and Immunity (2004) 5, 444–450. doi:10.1038/sj.gene.6364113 Published online 24 June 2004

Keywords: CCR5-D32; sclerosing cholangitis; lymphocyte trafficking

Introduction ized by a large number of plasma cells secreting IgG, which by binding complement contributes to tissue Inflammatory bowel disease (IBD) represents one of the damage and ongoing disease.4 Within IBD, there is a major chronic disorders of the gastrointestinal tract in subgroup of patients who develop primary sclerosing developed countries. The majority of patients are cholangitis (PSC), a chronic fibrosing liver disease diagnosed with either ulcerative colitis (UC) or Crohn’s characterized by progressive destruction of both intra- disease (CD) and recent epidemiological studies suggest hepatic and extrahepatic bile ducts by a combination of that the combined prevalence of these diseases may be as inflammation and fibrosis.5 A majority of PSC patients high as 0.5% of the general population.1 Both diseases are have UC (60–100%), whereas a much smaller number of associated with episodes of acute or chronic inflamma- UC patients develop the liver disease (2.4–7.5%) in tion affecting either the large bowel alone (UC) or both Caucasian populations.6 Both diseases share certain the small and the large bowel (CD).2 In CD, this intestinal autoimmune features including their close association inflammation is characterized by the presence of acti- with other autoimmune diseases and the presence of vated T-lymphocytes and macrophages, and disease both an antineutrophil cytoplasmic antibody (ANCA) pathogenesis is based on an inappropriate response to and antibody to tropomyosin.7 Other studies on the some commensal bacteria in a genetically susceptible mucosal immune system in UC patients have demon- host.3 In UC, the inflammatory cell infiltrate is character- strated large numbers of IgG plasma cells and elevated interleukin-5 (IL-5) levels, properties that support a humoral or T helper (Th)2 immune response.8 Correspondence: Dr G Radford-Smith, Department of Gastroenterology, The role of genetic susceptibility to CD has recently Level 9A, Ned Hanlon Building, Brisbane 4029, Australia. E-mail: [email protected] been strengthened by the successful identification of the Received 21 January 2004; revised 5 April 2004; accepted 5 April NOD2 gene on chromosome 16 using linkage analysis 2004; published online 24 June 2004 and positional cloning.9,10 In contrast, genetic studies in CCR5-D32 and primary sclerosing cholangitis REriet al 445 UC and PSC have predominantly used an association We have carried out a detailed analysis of the approach and focused on the HLA region on chromo- frequency of the CCR5-D32 mutation in an unselected some 6 and on key immunoregulatory genes including series of IBD patients from the Brisbane IBD Research tumour necrosis factor alpha (TNF-a).11,12 This may be in Group database at the Royal Brisbane Hospital (RBH) part due to a smaller number of affected sib-pairs in UC and a consecutive series of PSC patients from the IBD cohorts, a paucity of multiply affected PSC families, and database (RBH and Mater Hospital) and the Queensland the later age of presentation in PSC patients. Interest in Liver Transplant Service (Princess Alexandra Hospital). the genetics of PSC has been recently rekindled by two The study has investigated the role of this mutation in association analyses investigating the relevance of the disease susceptibility and in shaping the characteristics TNF-a and stromolysin genes in the development of this of both the intestinal and liver disease. disease.12,13 While the association with the TNF2 allele was considered secondary to the known association with the A1-B8-DRB1*0301-DQA1*0501-DQB1*0201 haplo- type, the association with the stromolysin gene was Results thought to be an independent one that may also Patient characteristics influence disease progression. However, the patients in Details of the major clinical characteristics of each patient this second study were not HLA typed.13 group are given in Table 1. There was complete clinical Recent interest in immunoregulatory genes has also information for all patients with UC or CD and for 65/71 focused on the chemokine family that are intimately patients with PSC. Over half of the PSC patients had involved in leucocyte trafficking and recruitment, and aggressive disease based on the need for liver transplan- therefore are likely to play a crucial role in determining tation (52%) and/or the presence of portal hypertension which cells migrate to the injured and/or inflamed (59%). Patients with PSC were significantly older (mean bowel and liver.14,15 Colonic epithelium shows upregula- age ¼ 51 years) than those with only UC (mean age ¼ 46 tion of RANTES (regulated upon activation normal T-cell years) or CD (mean age ¼ 41 years) (P ¼ 0.02 and expressed and secreted), which is the natural ligand for Po0.001, respectively). This was associated with a the chemokine receptor CCR5.16 During episodes of significantly longer duration of their underlying IBD intestinal inflammation, CCR5 is upregulated on activated Th1-type intestinal lamina propria and intrae- Table 1 Clinical details of the study patient population (given as pithelial lymphocytes as well as dendritic cells, mean7standard error or as a percentage of total in that group, supporting a role for this chemokine in lymphocyte where indicated) recruitment to the gut.17 CCR5 may also play a role in lymphocyte recruitment and immune responses within PSC UC CD the liver. This receptor, together with CXCR3, is (n ¼ 71) (n ¼ 162) (n ¼ 131) expressed at much higher levels on T cells isolated from the liver compared to the peripheral blood.18 Female/male (%)F 42/58 46/54 63/37 7 7 7 7 A 32-bp deletion in this gene (CCR5-D32 mutation) Mean age now (years s.e.)* 51 1.6 46 1.2 41 1.1 Mean age at diagnosis results in a nonfunctioning receptor that is trapped in the Of IBD 3071.8 3171.1 2971.2 endoplasmic reticulum and therefore not expressed at Of PSC 3771.8 the cell surface.19,20 The mutation is found in relatively high frequencies in the European population with an Mean disease duration** allele frequency of 10%.21 CCR5 is a coreceptor for the Of IBD 2171.4 1470.7 1370.6 7 M-tropic strain of human immunodeficiency virus (HIV) Of PSC 13 0.8 D and as a result CCR5- 32 homozygotes are afforded Location (%)+ complete immunity from this strain, while heterozygotes No IBD 12 show resistance to infection and a delay in the onset of Left-sided UC 1.5 45 acquired immunodeficiency syndrome (AIDS) compared Subtotal/total UC 77 55 to the wild-type counterparts.22–25 In chronic inflamma- Ileal/Ileocaecal CD 3 44 tory disorders, the CCR5-D32 mutation shows a negative Ileocolonic CD 1.5 39 26 Colonic CD 5 17 association with rheumatoid arthritis, which in turn is Immunosuppression (%)# 15 29 56 associated with a predominantly Th1 immune response. Transplant for PSC (%) 52 In patients with multiple sclerosis, the mutation is Severe PSC (%)++ 59 associated with a delayed age of onset of the disease Bowel resection (%) 29 35 60 and a lower risk of recurrent clinical disease activity.27,28 ## There have been six published papers on CCR5-D32 in Smoking status (%) Current smoker 10 2 33 IBD populations. One study found a significant associa- Ex-smoker 20 47 25 tion between CCR5-D32 homozygosity and perianal Never smoked 70 51 42 CD.29 There were no associations with disease susceptibility or other clinical features of IBD, and none of FGender: P ¼ 0.69 (UC vs PSC), P ¼ 0.008 (PSC vs CD). 30–34 these studies included PSC patients. PSC patients *Age now: P ¼ 0.02 (UC vs PSC), Po0.001 (PSC vs CD). have been included in two studies appearing in **IBD duration: P ¼ 0.0001 (UC vs PSC), Po0.0001 (PSC vs CD). 35,36 abstract form only. No association with CCR5-D32 +Location: P ¼ 0.02 (UC vs PSC). was found in the first analysis, but a significant positive #Immunosuppression: P ¼ 0.02 (UC vs PSC), Po0.0001 (PSC vs CD). association was reported in the second abstract.36 There ##Smoking: P ¼ 0.0005 (UC vs PSC), P ¼ 0.0004 (PSC vs CD). are no phenotypic details given in either of these ++Severe PSC as defined by the presence of portal hypertension abstracts. and/or need for liver transplantation.

Genes and Immunity CCR5-D32 and primary sclerosing cholangitis REriet al 446 (21 years (PSC) vs 14 years (UC) vs 13 years (CD), Table 3 CCR5-D32 allele frequencies in cases and controls P ¼ 0.0001 and Po0.0001, respectively). As expected, the majority of PSC patients had extensive colitis,37–39 which Patient population Number Wt D32 was significantly more common in PSC-UC compared to UC alone (P ¼ 0.02). Need for immunosuppression was Controls (%) 838 755 (90.1) 83 (9.9) less common for PSC-IBD patients (P ¼ 0.02). Smoking PSC* (%) 142 117 (82.4) 25 (17.6) UC (%) 324 284 (87.7) 40 (12.3) behaviour was significantly different between patients CD (%) 262 236 (90.1) 26 (9.9) with PSC and those with UC (P ¼ 0.0005) or CD IBD (UC+CD)** (%) 586 520 (88.7) 66 (11.3) (P ¼ 0.0004). Refractory UC (%) 164 142 (86.6) 22 (13.4) Extensive UC (%) 178 157 (88.2) 21 (11.8) PSC, with IBD (%) 122 102 (83.6) 20 (16.4) Allele frequencies and disease susceptibility PSC, no IBD (%) 20 15 (75) 5 (25) Genotyping for the CCR5-D32 mutation was successfully Severe PSC (%) 84 65 (77.4) 19 (22.6) performed on all patients and controls. None of the Mild PSC (%) 58 52 (89.7) 6 (10.3) patients and only two of the controls were homozygous for the mutated CCR5 gene, all other patients being *P ¼ 0.007, PSC vs controls. either heterozygous or homozygous for the wild-type **P ¼ 0.027, PSC vs IBD (UC+CD cases). allele. There was no evidence of deviation from Hardy– Weinberg equilibrium (HWE) in patients (P40.05 for all patient subgroups) or in controls (P ¼ 0.93). Comparison did not demonstrate any significant difference in the of those carrying a mutation vs wild-type individuals did distribution of the known PSC susceptibility alleles, HLA not show any significant differences in age of diagnosis, B8 (P ¼ 0.4), DR-2 (P ¼ 0.4), DR-3 (P ¼ 0.8) or DR-6 disease duration, sex or smoking status. Individuals (P ¼ 0.3), between the two groups (data not shown). heterozygous for the CCR5-D32 allele were significantly The CCR5-D32 allele frequency was similar in CD more common in the PSC group (35.2%) compared to patients (9.9%) to that found in controls (OR 1.16, controls (19.3%, OR 2.27, P ¼ 0.003, Table 2). Similarly, the P ¼ 0.59). Further subgroup analysis within the CD frequency of the CCR5-D32 allele was greater in patients population, including age at diagnosis, disease beha- with PSC (17.6%) compared to controls (9.9%, P ¼ 0.007) viour including the presence or absence of perianal and patients with IBD alone (11.3%, OR 1.9, P ¼ 0.027, disease, need for immunosuppression and need for Table 3). intestinal resection, did not reveal any association Patients with PSC but without underlying IBD (n ¼ 10, between the CCR5-D32 mutation and these character- mean disease duration 15.6 years) had a similarly istics (data not shown). elevated CCR5-D32 heterozygous genotype frequency (50%, Table 2) and allele frequency (25%, Table 3). Those CCR5 genotype, CCR5-D32 allele frequency and with UC alone did not show any significant differences disease characteristics in CCR5 genotype (OR 1.37, P ¼ 0.15, Table 2) or CCR5- This mutation was specifically associated with disease D32 allele frequency (P ¼ 0.23, Table 3) compared to severity in PSC with CCR5-D32 heterozygotes more controls. Complete HLA genotyping was available for frequent in the severe group (45%, Table 2) compared to the 37 PSC transplant patients. We therefore compared the rest (21%, OR 3.17, P ¼ 0.03). The CCR5-D32 allele the HLA allelic distribution among CCR5 wild-type frequency was similarly more common in this severe individuals vs those carrying a D32 mutation. Distin- group (23%) compared to the other PSC patients (10.3%, guishing wild-type and heterozygous subjects for CCR5 P ¼ 0.06). Analysis of CCR5 genotype based on transplantation alone showed a very similar distribution, with D32 significantly over-represented in transplant cases (46%) compared to nontransplant cases (24%, OR 2.76, Table 2 Genotype frequencies for CCR5-D32 in cases and controls P ¼ 0.05). There were five patients with portal hyperten- Patient population Number Wt/Wt Wt/D32 D32/D32 sion who did not undergo transplantation: two because of extensive cholangiocarcinoma and three because of age and significant comorbidities. These five patients Controls (%) 419 338 (80.5) 79 (19) 2 (0.5) PSC (%)* 71 46 (65) 25 (35) 0 were included in the severe group for all the above UC (%) 162 122 (75) 40 (25) 0 analyses. In all, 10 patients with PSC developed CD (%) 131 105 (80) 26 (20) 0 cholangiocarcinomas and their CCR5 genotype (40%, IBD (UC+CD) (%) 293 227 (77) 66 (23) 0 P ¼ 0.73) and CCR5-D32 allele frequency (20%, P ¼ 0.99) Refractory UCa (%) 82 60 (73) 22 (27) 0 b were similar to the PSC group as a whole. Extensive UC (%) 89 68 (76) 21 (24) 0 The CCR5-D32 allele frequency in subtotal or total UC PSC, with IBD (%) 61 41 (67) 20 (33) 0 PSC, no IBD (%) 10 5 (50) 5 (50) 0 was 11.8%, which was not significantly different from Severe PSCj (%) 42 23 (55) 19 (45) 0 those patients with distal disease (13.2%, P ¼ 0.71, Mild PSC (%) 29 23 (79) 6 (21) 0 Table 3) after adjusting for confounders. Thus, the higher frequency of CCR5-D32 in PSC-UC was not explained by *P ¼ 0.003, OR 2.27, PSC vs controls. disease extent. Nor was there a significant relationship jP ¼ 0.03, OR 3.17, severe PSC vs mild PSC—severe PSC defined as between severity of UC (as defined by colectomy or the presence of portal hypertension and/or the need for liver immunosuppression) and the presence of the CCR5-D32 transplantation. allele (P40.09). Similarly, there were no significant aRefractory defined as need for immunosuppression and/or differences in the CCR5 genotypes carried by these colectomy. subgroups of UC compared to controls. Patients with bExtensive UC defined as that extending beyond the splenic flexure. PSC had greater CCR5-D32 carriage rate than those with

Genes and Immunity CCR5-D32 and primary sclerosing cholangitis REriet al 447 UC but this did not reach statistical significance (P ¼ 0.1) a clue to the immune response associated with the CCR5- and diminished further after adjusting for potential D32 allele. These CCR5-deficient animals showed a confounders including age, disease duration, distribu- macrophage defect with significantly reduced levels of tion and smoking status (P ¼ 0.78). IL-1 and IL-6 compared to wild type, but normal levels of IL-10 and TNF-a. Further evidence for macrophage dysfunction was revealed by a significantly reduced hepatic clearance of Listeria infection again compared to Discussion wild-type mice. However, the CCR5-deficient mice Recent studies have confirmed the association between showed increased humoral responses to T-cell-depen- PSC and certain HLA haplotypes, as well as suggesting a dent antigenic challenge, a potentially interesting parallel role for polymorphisms within the matrix metalloprotei- to the humoral response in patients with PSC.47 nase family.12,13 This study provides evidence for an A ‘physiological’ inflammatory response already exists association between PSC and the chemokine receptor in the portal tracts of normal liver in the form of CCR5- family, with a significantly increased frequency of CCR5- positive lymphocytes.18 Ligands for CCR5, including D32 heterozygotes (P ¼ 0.003) and a significantly higher MIP-1a and MIP-1b, are constitutively expressed on CCR5-D32 allele frequency (P ¼ 0.007) in PSC patients portal vessels, thus allowing recruitment of CCR5- compared to matched controls. CCR5-D32 was also positive lymphocytes to these areas as part of an significantly commoner in PSC compared to a large, immunosurveillance role. This expression may be en- unselected series of patients with IBD alone (P ¼ 0.027). hanced during the development of inflammatory liver CCR5 represents a good candidate gene for patients diseases such as chronic hepatitis C and alcoholic with chronic liver disease in that it has an important role hepatitis,18,48 and may also be increased in both acute in the trafficking of activated mononuclear cells includ- and chronic liver allograft rejection.49 The results of the ing CD4 þ and CD8 þ memory T-lymphocytes to sites of current study suggest that the immunosurveillance role inflammation, by binding chemokines including MIP-1a played by CCR5-positive lymphocytes may be compro- (macrophage inflammatory protein-1a), MIP-1b and mised under special circumstances. These may include RANTES.14,15 Furthermore, CCR5 expression is preferen- reduced CCR5 expression, a potential increase in anti- tially associated with a Th1 subset of lymphocytes, which genic load within the portal vein associated with colonic secrete proinflammatory cytokines including TNF-a and inflammation and the presence of other modifying genes g-interferon.40,41 Carriage of at least one mutated copy of that push the response towards inflammation and the gene is associated with reduced CCR5 expression on fibrosis rather than resolution. the cell surface and may influence the strength and In this study, controls were matched for age and direction of the immune response in these individuals.19 ethnicity from a large, community-based twin registry. In This has been confirmed in animal studies including the addition, the control allele frequency was consistent with CCR5 knockout mouse that develops a Th2-driven previous studies from Caucasian populations.24 Impor- response in preference to a Th1 response, to a range of tantly, the PSC cohort in this study was particularly well inflammatory stimuli.42 The data presented in this study characterized. Relevant and detailed clinical information support a role for the CCR5-D32 mutation in PSC disease was available for over 90% of the patients and, susceptibility and disease severity. specifically, the diagnosis of PSC was based on both The pathogenesis of PSC remains unknown. However, radiological and histological records for all patients there are a number of studies that support a role for an included in the study. The inclusion of a large cohort of aberrant immune response in the pathogenesis of the both UC and CD patients in our study has also permitted disease including the close association with organ- investigation of the relationship between PSC and UC, specific autoimmune disorders, the presence of a number and examination of the effects of the CCR5-D32 mutation of autoantibodies in these patients and the significant on a Th1-mediated disorder, CD. link with HLA class I and class II genes. Importantly, the Carriage of the CCR5-D32 allele was greater for PSC association between PSC and some of these immune compared to UC, and both were higher than controls. disorders is significantly stronger than their association However, this only reached statistical significance for the with UC or IBD alone. This includes the links with PSC group and was specifically associated with severe autoimmune disorders,43 with anticolon antibodies44 and liver disease indicated by the presence of portal with HLA-B8 and DR03.45,46 The potential confounding hypertension or need for transplantation. This was effect of HLA genotype on the association between further supported by the results of the phenotype– CCR5-D32 and PSC was analysed in a subgroup of liver genotype analysis, with a higher CCR5-D32 carriage rate patients with severe disease and hence the strongest in severe PSC patients compared to the rest of this group association with the mutation. No significant difference (OR 3.17, P ¼ 0.03). Further analysis examining the effects was found in the distribution of PSC-related HLA of refractory or extensive UC with PSC did not genotypes (HLA-B8, DR-2, DR-3 and DR-6) in CCR5- demonstrate any significant association between the D32 heterozygotes compared to wild-type individuals. CCR5-D32 allele and the characteristics of the underlying The increased frequency of the CCR5-D32 allele that IBD. we have found in PSC patients compared to IBD alone or The strong association between CCR5-D32 and severe to controls may skew the immune response to recurrent liver disease in PSC is further evidence to suggest that or high levels of antigenic stimuli in the liver in a this disorder displays significant clinical and genetic direction that favours chronic, recurrent inflammation heterogeneity. The results support recent clinical obser- over resolution of the inflammatory response. This may vations on the natural history of PSC and genetic studies involve a switch to a Th2 response or some other that link certain HLA genotypes with severe, progressive mechanism. A murine model lacking CCR5 has provided disease.50,51 Importantly, our analysis of the PSC trans-

Genes and Immunity CCR5-D32 and primary sclerosing cholangitis REriet al 448 plant population carrying a high rate of CCR5-D32 on the results of both gastroscopy and abdominal indicated that any association with this mutation was contrast CT scan. independent of the HLA.51 Investigation of the CCR5- Patients with UC or CD were recruited from the D32 mutation in other forms of chronic liver disease is Brisbane IBD database as a consecutive series. The limited to hepatitis C and patients who have undergone database records the details of patients with IBD seen liver transplantation.52–55 In one of three studies on at the major public hospitals within Brisbane as part of a hepatitis C, the CCR5-D32 mutation was associated with clinical research initiative. PSC patients were similarly significantly less portal inflammation but greater fibro- recruited from this database and from the Queensland sis,52 while in the other two studies D32 was not liver transplant service. Nontransplant patients were associated with any specific outcomes in this popula- either from within Brisbane or referred from smaller tion.53,54 In contrast, analysis of CCR5-D32 in 146 patients units within Queensland to the RBH and Mater undergoing liver transplantation for multiple different hospitals, as tertiary institutions. Transplant patients indications demonstrated a significant association be- were similarly referred from within Brisbane and tween D32 and development of ischaemic-type biliary Queensland, with a smaller number from New Zealand. lesions postoperatively.55 All patients included in the study were Caucasian. The functional significance of the CCR5-D32 mutation In total, 419 healthy controls were obtained from has been extensively investigated both in homozygotes samples of the Australian twin registry collected for and heterozygotes, and in association with infectious other purposes. Only one twin from each monozygotic disease and inflammatory disease outcomes. Heterozy- pair was randomly selected for the control sample and gotes demonstrate reduced CCR5 expression, which is the twins were unselected for any disease phenotype. thought to be almost exclusively related to gene dosage They were matched on age and ethnicity with all PSC and not receptor sequestration.22 Nevertheless, the novel patients and a subset of the UC and CD patients. Ethical association also raises the possibility of the CCR5 gene approval was obtained from the research ethics commit- being in linkage disequilibrium with other gene(s). tee of the hospital (RBH) and from the research institute Despite this possibility, the CCR5 gene remains an where the genotyping was performed (Queensland attractive candidate gene for PSC in view of the well- Institute of Medical Research). characterized functional effects of the mutation and the role of the receptor in the immune response. Reduced CCR5 expression would reduce recruitment of lympho- DNA extraction cytes and other mononuclear cells to the liver. Patients Genomic DNA was extracted from EDTA-preserved with progressive PSC may mount an inappropriate whole blood using a salting out technique57 for all IBD response to certain environmental pathogens either and 50 PSC cases. DNA was extracted from formalin- acquired from the peripheral blood or from the gut, fixed, paraffin-embedded tissue sections from the re- which is aggravated by the presence of the CCR5-D32 maining 21 PSC patients using a previously described mutation. technique.58 In summary, we have found a significant association between the CCR5-D32 mutation and PSC, but not with either UC or CD. The results do not support a significant association with UC extent or severity but do indicate Genotyping that the mutation may influence PSC progression, with a Genomic DNA (100 ng per polymerase chain reaction significantly stronger association seen in patients with (PCR)) from each individual was amplified by an allele- severe liver disease compared to those with mild disease. specific PCR technique in a total volume of 21 ml, in buffer containing 100 mM Tris-Cl, pH 8.8, 50 mM KCl,

1.5 mM MgCl2, 0.2 mM of each dNTPs, 1 U Taq polymerase Materials and methods and 12.5 pmol of each primer. The primer sequences were as follows: sense, ATCACTTGGGTGGTGG Patients and controls CTGTGTTTGCGT CTC; antisense, AGTAGCAGAT In total, 162 patients with UC, 131 patients with CD and GACCATGACAAGCAGCGGCAG, corresponding to 71 patients with PSC were included in the study. The bases 505–535 and 667–697, respectively, of the published diagnosis of either UC or CD was based on standard sequence.59 Cycling conditions were as follows: 941C for clinical, radiological and histological criteria.56 The 5 min, followed by 30 cycles of 941C, 30 s; 701C, 30 s with diagnosis of PSC relied upon typical radiological and a 1 s increment per cycle, using an MJ research thermal histological appearances with all patients included in the cycler. The polymorphism was detected following study having at least an ERCP (endoscopic retrograde electrophoresis of the products in 10% polyacrylamide cholangiopancreatogram) or MRCP (magnetic resonance gels, staining with ethidium bromide and visualization cholangiopancreatogram) and a liver biopsy. on a UV transilluminator. Allele sizes were 193 bp for Disease severity in UC and CD was based on the need normal and 161 bp for the deletion allele. for immunosuppression to control the disease (at least 12 All samples were genotyped for the mutation on two continuous months of azathioprine, 6-mercaptopurine, separate occasions and there was complete concordance methotrexate or mycophenolate) or failure of medical for the two sets of results when analysed by two therapy requiring intestinal resection. In the PSC group, independent observers (RE and GR-S) who were blinded patients were divided into those with portal hyperten- to the clinical status of the patients at the time of gel sion and/or liver transplant (severe, progressive disease) analysis. In addition to the CCR5 genotyping, HLA or those without either of these parameters (mild genotypes were available for all the transplanted patients disease). The presence of portal hypertension was based (n ¼ 37).

Genes and Immunity CCR5-D32 and primary sclerosing cholangitis REriet al 449 Statistical analysis 14 Mackay CR. Chemokines: immunology’s high impact factors. HWE was assessed for patients and controls by compar- Nat Immunol 2001; 2: 95–101. ing the observed numbers of different genotypes with 15 Gerard C, Rollins BJ. Chemokines and disease. Nat Immunol those expected under HWE for the estimated allele 2001; 2: 108–115. frequency and by comparing Pearson’s goodness-of- 16 Dwinell MB, Eckmann L, Leopard JD, Varki NM, Kagnoff MF. statistic with a w2 distribution with one degree of Chemokine receptor expression by human intestinal epithelial cells. Gastroenterology 1999; 2: 359–367. freedom. The crude associations between CCR5 mutation 17 Agace WW, Roberts I, Wu L, Greinede C, Ebert EC, status and patient characteristics were assessed using the Parker CM. Human intestinal lamina propria and intraepithe- 2 Pearson’s w statistic, along with the relevant odds ratios lial lymphocytes express receptors specific for chemo- (and 95% confidence intervals) to quantify the size of kines induced by inflammation. Eur J Immunol 2000; 30: association. Logistic regression analysis was also con- 819–826. ducted to assess the significance of the associations 18 Shields PL, Morland CM, Salmon M, Qin S, Hubscher SG, between patient factors and CCR5 mutations, after Adams DH. Chemokine and chemokine receptor interactions adjusting for the potential confounding by age, duration provide a mechanism for selective T cell recruitment to of disease, location and smoking status. Odds ratios were specific liver compartments within hepatitis C-infected liver. calculated to estimate the relative risk of a particular J Immunol 1999; 163: 6236–6243. 19 Venkatesan S, Petrovic A, Van Ryk DI, Locati M, Weissman D, disease class in those people who were carriers of a Murphy PM. Reduced cell surface expression of CCR5 in mutated CCR5 allele. CCR5Delta 32 heterozygotes is mediated by gene dosage, rather than by receptor sequestration. J Biol Chem 2001; 277: 2287–2301. 20 Blanpain C, Lee B, Tackoen M et al. Multiple nonfunctional Acknowledgements alleles of CCR5 are frequent in various human populations. Blood 2000; 96: 1638–1645. Dr Eri was funded by a research grant from the 21 Stephens JC, Reich DE, Goldstein DB et al. Dating the origin of Queensland Cancer Fund (2000–2002). The Brisbane the CCR5-Delta32 AIDS-resistance allele by the coalescence of IBD Research Group’s database is supported by an haplotypes. Am J Hum Genet 1998; 62: 1507–1515. educational grant from Pfizer Australia. 22 Dean M, Carrington M, Winkler C et al. Genetic restriction of HIV-1 infection and progression to AIDS by a deletion allele of the CKR5 structural gene. Science 1996; 273: 1856–1862. 23 Huang Y, Paxton WA, Wolinsky SM et al. The role of a mutant References CCR5 allele in HIV-1 transmission and disease progression. Nat Med 1996; 3: 252–253. 1 Montgomery SM, Morris DL, Thompson NP, Subhani J, 24 Zimmerman PA, Buckler-White A, Alkhatib G et al. Inherited Pounder RE, Wakefield AJ. Prevalence of inflammatory bowel resistance to HIV-1 conferred by an inactivating mutation in disease in British 26 year olds: national longitudinal birth CC chemokine receptor 5: studies in populations with cohort. Br Med J 1998; 316: 1058–1059. contrasting clinical phenotypes, defined racial background, 2 Podolsky DK. Inflammatory bowel disease (1). N Engl J Med and quantified risk. Mol Med 1997; 3: 23–36. 1991; 325: 928–937. 25 Marmor M, Sheppard HW, Donnell D et al. Homozygous and 3 Shanahan F. Crohn’s disease. Lancet 2002; 359: 62–69. heterozygous CCR5-Delta32 genotypes are associated with 4 Thoree VC, Golby SJ, Boursier L et al. Related IgA1 and IgG- resistance to HIV infection. J Acquir Immune Defic Syndr 2001; producing cells in blood and diseased mucosa in ulcerative 27: 472–481. colitis. Gut 2002; 51: 44–50. 26 Gomez-Reino JJ, Pablos JL, Carreira PE et al. Association of 5 Chapman RW. Aetiology and natural history of primary rheumatoid arthritis with a functional chemokine receptor, sclerosing cholangitis: a decade of progress? Gut 1991; 32: CCR5. Arthritis Rheum 1999; 42: 989–992. 1433–1435. 27 Barcellos LF, Schito AM, Rimmler J et al. CC-chemokine 6 Broome U, Olsson R, Loof L et al. Natural history and receptor 5 polymorphism and age of onset in familial multiple prognostic factors in 305 Swedish patients with primary sclerosis. Immunogenetics 2000; 51: 281–288. sclerosing cholangitis. Gut 1990; 31: 961–963. 28 Sellebjerg F, Madsen HO, Jensen CV, Jensen J, Garred P. CCR5 7 Snook J. Are the inflammatory bowel diseases autoimmune delta32, matrix metalloproteinase-9 and disease activity in disorders? Gut 1990; 31: 961–963. multiple sclerosis. J Neuroimmunol 2000; 102: 98–106. 8 Radford-Smith GL. Ulcerative colitis—an immunological 29 Rector A, Vermeire S, Thoelen I et al. Analysis of the CC disease? Bailliers Clin Gastroenterol 1997; 11: 35–52. chemokine receptor 5 (CCR5) delta-32 polymorphism in 9 Hugot JP, Chamaillard M, Zouali H et al. Association of NOD2 inflammatory bowel disease. Hum Genet 2001; 180: 190–193. leucine-rich repeat variants with susceptibility to Crohn’s 30 Craggs A, Welfare M, Donaldson PT, Mansfield JC. The CC disease. Nature 2001; 411: 599–603. chemokine receptor 5 delta32 mutation is not associated with 10 Ogura Y, Bonen DK, Inohara N et al. A frameshift mutation in inflammatory bowel disease (IBD) in NE England. Genes NOD2 associated with susceptibility to Crohn’s disease. Immun 2001; 2: 114–116. Nature 2001; 411: 603–606. 31 Martin K, Heinzlmann M, Borchers R, Mack M, Loeschke K, 11 Satsangi J, Welsh KI, Bunce M et al. Contribution of genes of Folwaczny C. Delta 32 mutation of the chemokine-receptor 5 the major histocompatibility complex to susceptibility and gene in inflammatory bowel disease. Clin Immunol 2001; 98: disease phenotype in inflammatory bowel disease. Lancet 18–22. 1996; 347: 1212–1217. 32 Herfarth H, Pollok-Kopp B, Goke M, Press A, Oppermann M. 12 Mitchell SA, Grove J, Spurkland A et al. Association of Polymorphism of CC chemokine receptors CCR2 and CCR5 in the tumour necrosis factor alpha –308 but not the interleukin Crohn’s disease. Immunol Lett 2001; 77: 113–117. 10–627 promoter polymorphism with genetic susceptibility to 33 Hampe J, Lynch NJ, Daniels S et al. Fine mapping of the primary sclerosing cholangitis. Gut 2001; 49: 288–294. chromosome 3p susceptibility locus in inflammatory bowel 13 Satsangi J, Chapman RW, Haldar N et al. A functional disease. Gut 2001; 48: 191–197. polymorphism of the stromelysin gene (MMP-3) influences 34 Paavola P, Helio T, Kiuru M et al. Genetic analysis in Finnish susceptibility to primary sclerosing cholangitis. Gastroenterol- families with inflammatory bowel disease supports linkage to ogy 2001; 121: 124–130. chromosome 3p21. Eur J Hum Genet 2001; 9: 328–334.

Genes and Immunity CCR5-D32 and primary sclerosing cholangitis REriet al 450 35 Donaldson PT, Agarwal K, Saarinen S et al. Investigation of 48 Afford SC, Fisher NC, Neil DA et al. Distinct patterns of CCR5-D32 deletion and Fas/Apo-1 (CD95) single nucleotide chemokine expression are associated with leukocyte recruit- polymorphisms in primary sclerosing cholangitis (PSC). ment in alcoholic hepatitis and alcoholic cirrhosis. J Pathol Hepatology 2000; 32: 174A. 1998; 186: 82–89. 36 Satsangi J, Simmons J, Marshall S et al. CCR5-D32 polymorph- 49 Goddard S, Williams A, Morland C et al. Differential ism in inflammatory bowel disease: further association with expression of chemokines and chemokine receptors shapes ulcerative colitis and with primary sclerosing cholangitis. the inflammatory response in rejecting human liver trans- Gastroenterology 2000; 118: A337. plants. Transplantation 2001; 72: 1957–1967. 37 Fausa O, Schrumpf E, Elgjo K. Relationship of inflammatory 50 Bjornsson E, Boberg KM, Cullen S et al. Patients with small bowel disease and primary sclerosing cholangitis. Semin Liver duct primary sclerosing cholangitis have a favourable long Dis 1991; 11: 31–39. term prognosis. Gut 2002; 51: 731–735. 38 Boberg KM, Schrumpf E, Fausa O et al. Hepatobiliary disease 51 Boberg KM, Spurkland A, Rocca G et al. The HLA-DR3, DQ2 in ulcerative colitis. An analysis of 18 patients with hepato- heterozygous genotype is associated with an accelerated biliary lesions classified as small-duct primary sclerosing progression of primary sclerosing cholangitis. Scand J Gastro- cholangitis. Scand J Gastroenterol 1994; 29: 744–752. enterol 2001; 36: 886–890. 39 Sandborn WJ, Loftus Jr EV, Ahlquist DA. Association of 52 Hellier S, Frodsham AJ, Hennig BJ et al. Association of genetic primary sclerosing cholangitis and colorectal cancer in variants of the chemokine receptor CCR5 and its ligands, patients with ulcerative colitis: is it true and does it matter? RANTES and MCP-2, with outcome of HCV infection. Gastroenterology 1998; 115: 236–237. Hepatology 2003; 38: 1468–1476. 40 Loetscher P, Uguccioni M, Bordoli L et al. CCR5 is character- 53 Promrat K, McDermott DH, Gonzalez CM et al. Association of istic of Th1 lymphocytes. Nature 1998; 391: 344–345. chemokine system polymorphisms with clinical outcomes and 41 Qin S, Rottman JB, Myers P et al. The chemokine receptors treatment responses of chronic hepatitis C. Gastroenterology CXCR3 and CCR5 mark subsets of T cells associated with 2003; 124: 352–360. certain inflammatory reactions. J Clin Invest 1998; 101: 746–754. 54 Wasmuth HE, Werth A, Mueller T et al. CC chemokine 42 Andres PG, Beck PL, Mizoguchi A et al. Mice with a selective receptor 5 Delta32 polymorphism in two independent cohorts deletion of the CC chemokine receptors 5 and 2 are protected of hepatitis C virus infected patients without hemophilia. from dextran sodium sulfate-mediated colitis: lack of CC J Mol Med 2004; 82: 64–69. chemokine receptor 5 expression results in an NK1.1+lym- 55 Moench C, Uhrig A, Lohse AW, Otto G. CC chemokine phocyte-associated Th2 type immune response in the intes- receptor 5 Delta32 polymorphism—a risk factor for ischaemic- tine. J Immunol 2000; 164: 6303–6312. type biliary lesions following orthotopic liver transplantation. 43 Saarinen S, Olerup O, Broome U. Increased frequencies of Liver Transplant 2004; 10: 434–439. autoimmune diseases in patients with with primary sclerosing 56 Lennard-Jones JE. Classification of inflammatory bowel cholangitis. Am J Gastroenterol 2000; 95: 3195–3199. disease. Scand J Gastroenterol Suppl 1989; 170: 2–6. 44 Chapman RW, Cottone M, Selby WS, Shepherd HA, Sherlock 57 Miller SA, Dykes DD, Polesky HF. A simple salting out S, Jewell DP. Serum autoantibodies, ulcerative colitis and procedure for extracting DNA from human nucleated cells. primary sclerosing cholangitis. Gut 1986; 27: 86–91. Nucleic Acids Res 1988; 16: 1215. 45 Schrumpf E, Gjone E. Hepatobiliary disease in ulcerative 58 Levi S, Urbano-Ispizua A, Gill R et al. Multiple K-ras codon 12 colitis. Scand J Gastroenterol 1982; 17: 961–964. mutations in cholangiocarcinomas demonstrated with a 46 Chapman RW, Varghese Z, Gaul R, Patel G, Kokinon N, sensitive polymerase chain reaction technique. Cancer Res Sherlock S. Association of primary sclerosing cholangitis with 1991; 51: 3497–3502. HLA-B8. Gut 1983; 24: 38–41. 59 Raport CJ, Gosling J, Schweickart VL, Gray PW, Charo IF. 47 Zhou Y, Kurihara T, Ryseck RP et al. Impaired macrophage Molecular cloning and functional characterization of a function and enhanced T cell-dependent immune response in novel human CC chemokine receptor (CCR5) for RANTES, mice lacking CCR5, the mouse homologue of the major HIV-1 MIP-1beta, and MIP-1alpha. J Biol Chem 1996; 271: coreceptor. J Immunol 1998; 160: 4018–4025. 17161–17166.

Genes and Immunity