BRIP1 coding variants are associated with a high risk of hepatocellular carcinoma occurrence in patients with HCV- or HBV-related liver disease

The molecular mechanisms of hepatocellular carcinoma (HCC) carcinogenesis are still not fully understood. DNA repair defects may influence HCC risk. The aim of the study was to look for potential genetic variants of DNA repair genes associated with HCC risk among patients with alcohol- or viral-induced liver disease. We performed four case-control studies on 2,006 European- (Derivation#1 and #2 studies) and African-ancestry (Validation#1 and #2 studies) patients originating from several cohorts in order to assess the association between genetic variants on DNA repair genes and HCC risk using a custom array encompassing 94 genes. In the Derivation#1 study, the BRIP1 locus reached array-wide significance (Chi-squared SV-Perm, P=5.00×10−4) among the 253 haplotype blocks tested for their association with HCC risk, in patients with viral cirrhosis but not among those with alcoholic cirrhosis. The BRIP1 haplotype block included three exonic variants (rs4986763, rs4986764, rs4986765). The BRIP1 ‘AAA’ haplotype was significantly associated with an increased HCC risk [odds ratio (OR), 2.01 (1.19–3.39); false discovery rate (FDR)-P=1.31×10−2]. In the Derivation#2 study, results were confirmed for the BRIP1 ‘GGG’ haplotype [OR, 0.53 (0.36–0.79); FDR-P=3.90×10−3]. In both Validation#1 and #2 studies, BRIP1 ‘AAA’ haplotype was significantly associated with an increased risk of HCC [OR, 1.71 (1.09–2.68); FDR-P=7.30×10−2; and OR, 6.45 (4.17–9.99); FDR-P=2.33×10−19, respectively]. Association between the BRIP1 locus and HCC risk suggests that impaired DNA mismatch repair might play a role in liver carcinogenesis, among patients with HCV- or HBV-related liver disease.


Description of derivation (Derivation #1 and #2) and validation studies (Validation #1 and #2) Derivation #1 study
This investigation is an ancillary study of the CiRCE multicenter case-control study, which aimed to assess the influence of metabolic, genetic, drugs, and environmental factors (alcohol, tobacco, viruses, and diet) on the risk of HCC among cirrhotic patients (1,2). Since November 1 st , 2008, cirrhotic European-ancestry patients with or without HCC were included in six university hospitals belonging to the Cancéropôle Grand-Est (CGE) consortium (Besançon, Dijon, Metz, Nancy, Reims, and Strasbourg). All participants gave their written informed consent and the protocol of the study was approved by the ethic committee of the University Hospital of Dijon (University of Burgundy, Dijon, France) who coordinated the CiRCE study research protocol at the national level (French National Research Agency, ANR-11-LABX-0021) (1,2). Each study center followed the same study protocol (1,2). Cases were defined as cirrhotic patients presenting with HCC. The diagnosis of cirrhosis was based on liver biopsy or, in the absence of liver biopsy, on typical clinical, morphological and biological data at the discretion of the physician according to international guidelines. HCC diagnosis was based on the European Association for the Study of the Liver (EASL) criteria (3). The absence of HCC in patients with liver cirrhosis at inclusion was assessed through high-quality imaging examinations (abdominal ultrasonography, abdominal computed tomography scan, or abdominal magnetic resonance imaging) and alpha-fetoprotein <100 ng/ml within the 2 months prior to inclusion (2). The diagnosis of alcoholic cirrhosis was based on an alcohol consumption >14 units/week for women and >28 units/week for men or a report of current or previous alcohol abuse (2). Exclusion criteria were as follows: patients under 35 years old in order to avoid the inclusion of patients whose cancers have genetic basis since this specific population is not representative of the large majority of patients thus introducing an analytical bias; progressive extrahepatic cancer; human immunodeficiency virus infection; acute alcoholic hepatitis; major somatic or psychiatric illness not compatible with inclusion in the CiRCE study; or non-HCC primary liver cancer (1,2). Age, sex, etiology of cirrhosis (HCV or HBV), alcohol, and tobacco consumption were obtained through standardized questionnaires. Blood samples collected from cirrhotic patients were processed at each study center. Aliquots of frozen white blood cells were shipped on dry ice to the INSERM unit U954 laboratory "NGERE -Nutrition, Genetics, and Environmental Risk Exposure" for DNA extraction and genomic analyses. An overview of the study design is reported in Figure 1A.

Derivation #2 study
In the Derivation #2 study the 56 HCC patients with viral cirrhosis from the Derivation #1 study were compared with 970 HCC-free and cirrhosis-free patients with chronic HCV infection. These patients were recruited from two cohorts of adult patients of European descent from France and Switzerland. As previously described, the French cohort (ANRS Genoscan study group, n=398) included patients from the hepatology units of several hospitals in Paris and Marseilles (4). The Swiss Hepatitis C Cohort Study (SCCS, n=572) is a multicenter study of HCV-infected patients enrolled at eight major Swiss hospitals and the affiliated local centers. The French and Swiss cohorts were genotyped for ~350,000 variants and ~1,000,000 variants, respectively, using Illumina HumanCNV370-Duo and Human1M-Duo beadchips (Illumina, San Diego, USA) (4). Genotype imputation was performed in the French cohort (for the rs4986764 variant), using the Swiss cohort as a template. Only genotyping data for the three top variants retrieved in the Derivation #1 study (BRIP1 locus; rs4986763, rs4986764, rs4986765) were used in the Derivation #2 study.

Validation #1 study
Two case-control comparisons were performed in the validation study. The Validation #1 study included patients who were consecutively referred to the Jean Verdier Hospital Liver Unit (JVH cohort) for diagnosis and management of cirrhosis between January 1999 and December 2007 (5). In order to replicate our initial results in African populations, only patients with African ancestry were included in the genotyping study. In the Validation #1 study, 136 HCC patients with viral-related cirrhosis from the JVH cohort were compared with 99 HCC-free patients with HBV-or HCV-related cirrhosis from the JVH cohort. All participants gave their informed consent. DNA samples were prepared from frozen blood samples stored in the Liver Biobank "CRB des Hôpitaux Universitaires Paris-Seine-Saint-Denis" BB-0033-00027, and were shipped on dry ice to the INSERM unit U954 laboratory for genomic analyses.

Validation #2 study
In the Validation #2 study, 136 HCC patients with viral-related cirrhosis from the JVH cohort were compared with 305 HCC-free and cirrhosis-free patients with HBV-and/or HCV infection recruited in Benin and Togo (Benin-Togo cohort) (6-10). Institutional review board approval was obtained from the ethical committees of the University Hospital of Nancy (Vandoeuvre-lès-Nancy, France), the University of Benin (Cotonou, Benin), and the University of Lomé (Lomé, Togo) (6-10). Written informed consent was obtained from participants. nlm.nih.gov; and Ensembl, http://www.ensembl.org/ index.html) based on evidence for their relationship with several malignancies, including HCC. We included genetic variants encoding non-synonymous amino acid substitutions that were suggested to alter protein function or those located within the 5' or 3' untranslated region of the gene that could potentially alter mRNA stability. Additionally, we included intronic variants in some instances to allow good genes coverage. We conducted a preliminary in silico validation phase of all identified genetic variants in order to determine their suitability for genotyping with the GoldenGate assay using the Illumina Assay Design Tool (ADT) at http://www. Illumina.com. The assessment of each genetic variant was conducted based on two Illumina's in-house criteria: 1) the designability score: 1, highly designable; 0.5, moderately designable; or 0, low designability; 2) the 60bp limitation rule (a genetic variant cannot be closer than 60-bp to another one on the oligonucleotide pool assay, OPA). Genetic variants with low designability score were discarded and the final selection of the 384 SNPs was sent to Illumina for the synthesis of the OPA (also referred to as GoldenGate genotyping assay). Full description of the 'DNA repair genes' custom array is summarized in the Supplemental Table S1 (See supplementary appendix).

Genetic variant selection for the 'DNA repair genes' custom array
The
The assay was performed on a LightCycler ® 480 (Roche Applied Science) under the following conditions: pre-incubation at 95 °C for 10 min each, followed by 45 cycles of 95 °C for 10 s (denaturation) and 60 °C for 15 s (annealing with ramp rate of 2.20 °C/s and 1 step per cycle down to 50 °C), and extension step of 72 °C 10 s. Acquisition of the fluorescence signal was performed during each extension step. Amplification was followed by HRM analysis of the real-time PCR products, consisting of a denaturation step at 95 °C for 60 s, a cooling step at 40 °C for 60 s, a short hold at 65 °C for 1 s and a continuous acquisition step (with 25 acquisitions per °C) from 65 °C to 95 °C. The final cooling step consisted of a 10 s hold at 40 °C. HRM curves were classified into two or three distinct groups. Samples with known genotypes were used as internal references to generate standard curves for the classification of the unknown samples.