Mutation analysis of the Fanconi anaemia A gene in breast tumours with loss of heterozygosity at 16q24.3

The recently identified Fanconi anaemia A (FAA) gene is located on chromosomal band 16q24.3 within a region that has been frequently reported to show loss of heterozygosity (LOH) in breast cancer. FAA mutation analysis of 19 breast tumours with specific LOH at 16q24.3 was performed. Single-stranded conformational polymorphism (SSCP) analysis on cDNA and genomic DNA, and Southern blotting failed to identify any tumour-specific mutations. Five polymorphisms were identified, but frequencies of occurrence did not deviate from those in a normal control population. Therefore, the FAA gene is not the gene targeted by LOH at 16q24.3 in breast cancer. Another tumour suppressor gene in this chromosomal region remains to be identified. © 1999 Cancer Research Campaign


Tumour material
Freshly frozen breast tumour tissue was collected at the Department of Pathology of the Leiden University Medical Center. Haematoxylin and eosin-stained sections of tumour tissue blocks were examined by a pathologist to select cases that contained at least 50% tumour cells. Tumours for FAA mutation analysis were selected by testing polymorphic markers on chromosome 16 as has been described previously (Cleton-Jansen et al, 1994). Seventeen tumours showing LOH only on chromosomal band 16q24.3, two tumours showing complex LOH on 16q22 and 16q24.3 but retention of markers in between and three controls without 16q LOH were used in this study. The LOH pattern on 16q of the tumours was shown by dense LOH mapping, which was described in detail by Moerland et al (1997). Genomic DNA from 50 healthy controls originating from the Dutch population was used for testing frequencies of polymorphisms.

Mutation screening
RNA was extracted from 10Ð20 sections of 20 µm from tumour tissue blocks using TRIzol reagent (GIBCO BRL) according to the manufacturerÕs instructions. cDNA was synthesized from 1 to 10 µg RNA in a 20 µl reaction volume containing 0.2 units AMV reverse transcriptase (RT; Boehringer), 1 µg random primed hexamers, 2 mM dNTP in the buffer supplied with the enzyme.

Mutation analysis of the Fanconi anaemia A gene in breast tumours with loss of heterozygosity at 16q24.3
Polymerase chain reaction (PCR) reactions were performed on 2 µl cDNA with primers derived from the FAA cDNA sequence. Table 1 shows the PCR strategy. Primer sequences and PCR conditions are available on request. PCR fragments FAA1 to FAA6 were radiolabeled by the addition of 2 µCi α-32 P-dCTP and digested with restriction enzymes resulting in fragments of appropriate length to be analysed by single-strand conformational polymorphism (SSCP). FAA7 and FAA8 were amplified in a first step, followed by a nested PCR, which resulted in 3 and 6 radiolabelled fragments, respectively, with a size range between 200 and 400 bp.
RT-PCR products were analysed on SSCP gels containing 10% glycerol, 6% acrylamide and TBE or MDE gels (Boehringer). If a variant band pattern was observed in an RT-PCR fragment, sequencing was performed using a Cycle Sequencing kit (Perkin Elmer). Subsequently, restriction enzyme analysis on PCR products amplified from genomic DNA of tumour and normal tissue of all patients and on DNA from healthy controls was applied to determine the frequency of the variant alleles (Table 2).
Genomic DNA from eight tumours was also tested in an exonby-exon SSCP analysis described elsewhere (Wijker et al, 1998). The 43 exons of FAA (Lanzano et al, 1997) were amplified from genomic DNA with primers located 40Ð60 bp from the exon boundaries with T7 (sense) and Sp6 (antisense) sequences to facilitate sequencing. Fragments were analysed by SSCP on a 20% acrylamide gel with silver staining to visualize the bands. Southern analysis to identify genomic rearrangements or large deletions was performed as described previously (Devilee et al, 1991). Table 1 shows the results of the SSCP analysis on cDNA of the tumours. No tumour-specific variations were found, either in cDNA or genomic DNA. SSCP analysis sometimes resulted in altered banding patterns. Upon sequencing of the PCR fragments showing variations, these always appeared to be polymorphisms of the FAA gene that could also be identified in a normal control population. Moreover, analysis of DNA from non-neoplastic cells of the same patient showed that these variants were also present in the germline and were, therefore, not tumour-specific. Both alleles of a polymorphism could be recognized in DNA from tumour tissue, due to a 20Ð50% contamination with non-neoplastic cells. Table 2 shows these polymorphisms, their location, nature, result on the amino acid sequence, and occurrence in breast tumours and in controls. Although all four polymorphisms resulted in an amino acid change, none of these appeared to be more predominant in patients with breast cancer and cannot, therefore, be considered as pathogenic germline mutations. An exception is the 17T/A (V/D6) variant, which was found in only 1 out of 100 chromosomes of the control population but in 2 out of 16 chromosomes in patients. However, in one patient, case BT555, there was LOH of the 17T allele, whereas in the other, case BT757, the 17A allele showed LOH, indicating that there is no preferential LOH for one allele or the other. This polymorphism was also identified in a Fanconi anaemia A family, but did not segregate with the disease phenotype . The V/D6 variant was only detected by Tumours with LOH only on 16q24

RESULTS
Tumours with complex LOH only on 16q24 and 16q22 a G indicates that this sample was also analysed in a genomic DNA based exon-by-exon SSCP. b N indicates that the sample was tested, but no variant bands were found. c Numbers refer to the position of a variant nucleotide in the cDNA sequence (1 = first basepair of the FAA start codon); letters indicate which alleles are identified in the sample. d Empty cells: sample not tested for this fragment.
SSCP analysis on genomic DNA, not on cDNA, most probably because the PCR primer for the genomic assay was located 50 bp upstream of the start codon, and the cDNA primer was only 7 bp upstream.
Evidence for large genomic deletions in FA patients has been reported (Fanconi Anemia/Breast Cancer Consortium, 1996). Since large deletions and other rearrangements cannot be detected by PCR a subset of seven tumours with LOH at 16q24.3, six breast tumour cell lines and two normal mammary epithelial cell lines, were examined by Southern analysis. DNA from breast tumour cell lines was obtained from CAMA-1, MCF7, MDA468, BT474, ZR75-1 and MPE600 (kindly provided by Dr Joe Gray), and from SV40 transformed mammary epithelial cell lines HBL100 and RC6. Four of these breast tumour cell lines show homozygosity for four highly polymorphic markers near the FAA gene, suggesting that LOH had occurred at 16q24. DNA was digested with MspI, TaqIor PvvuII, blotted onto Hybond N+ membranes and hybridized with an FAA full-length cDNA probe (provided by Dr Hans Joenje). None of the tumour samples or cell lines showed aberrant bands. The FAA cDNA recognized a variant band of 3.4 kb in Pvull digests of two out of seven tumours tested, but this was also present in non-neoplastic DNA and is, therefore, considered to be a polymorphism. Southern analysis of DNA from 11 healthy controls showed that this polymorphism could be found in two individuals, a frequency similar to that in patients.

DISCUSSION
The most distal region of chromosome arm 16q has been frequently implicated as a target for LOH in both breast cancer and prostate cancer and is, therefore, likely to contain a tumour suppressor gene. LOH studies of this chromosome arm in breast tumours have also implicated other candidate regions. The Ecadherin gene located at 16q22.1 was actually shown to be the target of LOH, i.e. one copy was affected by a protein truncating mutation and the other copy removed by LOH (Berx et al, 1995(Berx et al, , 1996. The E-cadherin gene, however, is only targeted in lobular breast tumours, which comprise a minority of 5Ð10% of the total number of cases. The Fanconi anaemia A susceptibility gene is located at 16q24.3 (Pronk et al, 1995;Gschwend et al, 1996) in the smallest region of overlap that was defined by detailed deletion mapping of 79 breast tumours (Cleton-Jansen et al, 1994). FAA encodes a unique protein that does not exhibit any homology to known proteins that might suggest a function. Recently, it was suggested that FA is due to a defective caretaker gene, as are ataxia-telangiectasia, Bloom syndrome, hereditary non-polyposis colorectal cancer, Werner syndrome and the nucleotide excision repair syndromes xeroderma pigmentosum, Cockayne syndrome and trichothiodystrophy . Each of these genes is responsible in a unique way for the integrity of the genome, and when mutated causes predisposition to cancer. FA patients are known to be at high risk for malignancy (Alter, 1996), while the cancer risk in FA carriers has not been well studied as yet. Cultured FA cells show baseline chromosomal instability as well as increased levels of chromosomal aberrations induced by DNA cross-linking agents. Since tumour cells are characterized by chromosomal instability, FAA was postulated to be a candidate for the gene targeted by LOH of this region. In addition, the breast tumour suppressor genes BRCA1 and BRCA2 were shown to play a role in DNA repair (Scully et al, 1997;Patel et al, 1998).
This report shows that, although a specific set of 19 breast tumours with isolated LOH of 16q24.3 was screened for mutations in the FAA gene using three different approaches, SSCP on cDNA, SSCP on genomic DNA and Southern analysis, no tumour-specific mutations were found. The SSCP analysis on genomic DNA was identical to that used in a study on FA patients (Wijker et al, 1998). This study shows that the method is able to identify mutations and polymorphisms. Moreover, our analysis is corroborated by the identification of naturally occurring polymorphisms in the FAA coding sequencing. The high number of polymorphisms found is in concordance with the recently reported variability of the FAA gene .
The FAA gene is rich in Alu sequences (Ianzano et al, 1997) suggesting Alu-mediated recombination resulting in large genomic deletions might be an important mechanism for the generation of FAA mutations. Several such genomic deletions, which would be missed by SSCP analysis, have been observed in FA patients (Fanconi Anemia/Breast Cancer Consortium, 1996;Centra et al, submitted;Levran et al, submitted). Therefore, a subset of seven tumours and six breast tumour cell lines was tested by Southern analysis to identify large genomic rearrangements, but this did not result in the identification of such mutations in breast tumours.
In conclusion, the FAA gene has been excluded as the breast tumour suppressor gene on 16q24.3 targeted by LOH. Therefore, another gene in this region remains to be identified.