For CTCF analysis, 153 breast cancer patients were recruited from 139 families. In our previous genome-wide linkage analysis, 28 of the families were shown to share a common haplotype in 16q22–24, where the CTCF gene resides. All the families included were recruited through a clinicogenetic counselling procedure and were considered BRCA1 and BRCA2 negative 1516. In total, 90 cases were ascertained from high-risk families in which there were three or more first-degree affected relatives with breast cancer over at least two generations. The mean number of cases in each family was 3.3 and the mean age of diagnosis was 53 years of age. The other 63 cases were ascertained from low-risk families in which there were only two first-degree affected relatives with breast cancer, with a mean age of onset of 47 years. All the cases in this study were included in accordance with guidelines approved by the Ethics Committee at Karolinska Institutet. As control population we used 190 unrelated healthy relatives of patients recruited at Department of Clinical Genetics, Karolinska Hospital.

DHPLC analysis was performed with automated instrumentation equipped with a DNASep column, the Wave^® nucleic acid fragment analysis system (Transgenomic, Santa Clara, CA), on the reverse transcriptase polymerase chain reaction (RT–PCR) product. The entire coding sequence of the gene CTCF was divided into six overlapping fragments; the primers used for RT–PCR amplification of each segment are given in Table 1. Total RNA of each sample was extracted from EBV-transformed lymphocytes using TRIzol^® Total RNA Isolation Reagent kit (Invitrogen, Carlsbad, CA) and was reverse transcribed with random hexamers using the GeneAmp^® RNA PCR kit (PE Biosystems, Foster City, CA) to generate cDNA. A 2 μl aliquot of cDNA was used in PCR amplification in 50 μl reaction volumes containing 10 pmol of sense and antisense primers for each fragment, 1.5 mM MgCl₂, 100 μM dNTPs, 1.25 U of AmpliTaq Gold polymerase (PE Biosystems) and 1 × PCR buffer supplied by the manufacturer. A universal Touchdown PCR protocol was employed to improve PCR specificity and to minimize PCR optimization. It consisted of an initial incubation of 95°C for 10 min, a first round of 6 cycles of 95°C for 30 s, 58°C for 45 s with 1°C decrement per cycle, 72°C for 45 s, a second round of 26 cycles of 95°C for 30 s, 53°C for 45 s, 72°C for 45 s, and a final extension step of 72°C for 7 min. PCR products were then denatured at 95°C for 5 min and cooled slowly in a PCR instrument at a rate of 1.5°C per cycle for 40 cycles. Each PCR product (10 μl) was then loaded on the Wave^® instrument and eluted with a linear acetonitrile gradient consisting of buffer A (0.1 M triethylamine acetate; TEAA) and buffer B (0.1 M TEAA in 25% acetonitrile) at a constant flow rate of 0.9 ml/min. Specific values of the gradient ranges and the column temperature required for optimal resolution of each amplicon were determined by the WaveMaker software (Transgenomic) based on the sequence. A Mutation Standard (Transgenomic) was run with the samples analysed, to ensure the best performance of the DHPLC instrument for mutation detection. The elution profiles were recorded and analysed by HSM 7000 software (Transgenomic).

Samples with an altered DHPLC profiles were reamplified and purified with QIAEX II gel extraction kit (Qiagen, Hilden, Germany). The bi-directional sequencing reaction was performed with ABI PRISM^® BigDye™ Terminator Cycle Sequencing Ready Reaction Kits version 2.0 (PE Biosystems) in accordance with the manufacturer's instructions, with the primers used in RT–PCR amplification as sequencing primers (Table 1).

Pyrosequencing was adopted in this study to determine the frequency of the CTCF variants found in familial breast cancer cases in normal controls. Primers used in the pyrosequencing analysis are listed in Table 2. The conditions of PCR amplification were the same as those used in the DHPLC analysis except that genomic DNA rather than cDNA was used as template and the number of cycles was increased from 35 to 50. Biotin-labelled amplicons (30 μl) were mixed with 25 μl of BW buffer (10 mM Tris-HCl, 2 M NaCl, 1 mM EDTA and 0.1 Tween 20, pH 7.6) and immobilized on 20 μl of streptavidin-coated super paramagnetic beads (Dynabeads^®, M-280–streptavidin; Dynal AS, Oslo, Norway) by incubation at 65°C for 15 min, with shaking. Single-stranded DNA was obtained by incubating the immobilized amplicons in 50 μl of 0.5 M NaOH for 5 min using a PSQ 96 Sample Prep Tool (Pyrosequencing AB, Uppsala, Sweden). Each sample (well) was washed twice with 100 μl of 1 × annealing buffer (200 mM Tris acetate and 50 mM magnesium acetate). The immobilized strand was resuspended in 45 μl of 1 × annealing buffer containing 2 pmol of sequencing primer. Hybridization was performed by incubation at 80°C for 2 min, followed by cooling to room temperature. Finally, real-time pyrosequencing was performed on an automated 96-well pyrosequencer instrument with a PSQ SNP Reagent Kit (including all required enzymes and substrates) provided by the manufacturer (Pyrosequencing AB). The base sequence interpretation of the chromatograms and SNP genotype recognition were implemented automatically by the software and checked manually.