Mutation and association analysis of GEN1 in breast cancer susceptibility

Abstract

GEN1 was recently identified as a key Holliday junction resolvase involved in homologous recombination. Somatic truncating GEN1 mutations have been reported in two breast cancers. Together these data led to the proposition that GEN1 is a breast cancer predisposition gene. In this article we have formally investigated this hypothesis. We performed full-gene mutational analysis of GEN1 in 176 BRCA1/2-negative familial breast cancer samples and 159 controls. We genotyped six SNPs tagging the 30 common variants in the transcribed region of GEN1 in 3,750 breast cancer cases and 4,907 controls. Mutation analysis revealed one truncating variant, c.2515_2519del-AAGTT, which was present in 4% of cases and 4% of controls. We identified control individuals homozygous for the deletion, demonstrating that the last 69 amino acids of GEN1 are dispensable for its function. We identified 17 other variants, but their frequency did not significantly differ between cases and controls. Analysis of 3,750 breast cancer cases and 4,907 controls demonstrated no evidence of significant association with breast cancer for six SNPs tagging the 30 common GEN1 variants. These data indicate that although it also plays a key role in double-strand DNA break repair, GEN1 does not make an appreciable contribution to breast cancer susceptibility by acting as a high- or intermediate-penetrance breast cancer predisposition gene like BRCA1, BRCA2, CHEK2, ATM, BRIP1 and PALB2 and that common GEN1 variants do not act as low-penetrance susceptibility alleles analogous to SNPs in FGFR2. Furthermore, our analyses demonstrate the importance of undertaking appropriate genetic investigations, typically full gene screening in cases and controls together with large-scale case–control association analyses, to evaluate the contribution of genes to cancer susceptibility.

Introduction

Breast cancer is twice as common in women with an affected first degree relative and germline mutations in the known breast cancer predisposition genes account for <30% of this excess familial risk. Inactivating mutations in BRCA1 and BRCA2 are high-penetrance breast cancer susceptibility alleles accounting for ~16%, whilst mutations in the functionally related DNA repair genes CHEK2, PALB2, ATM and BRIP1 are of intermediate penetrance, and account for <3% [1-7]. Genome-wide association studies have identified 18 common variants which have been classed as low-penetrance breast cancer predisposition alleles. When combined these SNPs account for approximately 8% of familial disease risk [8-15].

The breast cancer predisposition genes BRCA1, BRCA2, CHEK2, ATM, BRIP1 and PALB2 are involved in double-strand break repair via the homologous recombination pathway [16-18]. This pathway repairs breaks caused by ionising radiation and mutagenic chemicals by utilising the homologous chromosome as a template for repair [19]. During this process a covalent link between each pair of homologous chromatids is formed and is known as a Holliday junction. Once repair is complete, these junctions are resolved by symmetrical nicking of the DNA strands, followed by separation and ligation to form two separate duplex molecules. Holliday junction resolvases mediate the transition from four covalently bonded chromatids to two separate duplex chromosomes [20]. In 2008, GEN1 was identified as the gene encoding a human Holliday junction resolvase with a key a role in this recombinational repair pathway [21].

Through genome-wide sequencing of the exome in breast cancer cell lines and primary tumours, two somatic frameshift mutations in GEN1 were identified [22, 23]. This, together with recognition of the role of GEN1 in DNA repair, led to the conclusion that constitutional GEN1 mutations would confer susceptibility to breast cancer in a fashion analogous to some other DNA repair genes [21]. However, to date, no data to support this conclusion have been published. In order to investigate formally the contribution of GEN1 to breast cancer susceptibility, we have undertaken mutational analysis of the full gene in 192 breast cancer cases and 184 controls and an association analysis of common variants in the vicinity of GEN1 in constitutional DNA from 3,750 breast cancer cases and 4,907 controls.

Materials and methods

Samples

Cases were unrelated individuals with breast cancer and a family history of breast cancer that were recruited through cancer genetics clinics in the UK, through the Genetics of Familial Breast Cancer Study. Informed consent was obtained from all family members and the research was approved by the London Multicentre Research Ethics Committee (MREC/01/2/18). Samples from non-Caucasian UK ethnic groups were excluded. The extent of family history was quantified using a Family History Score, defined by the number of relatives with breast cancer, weighted by their degree of relatedness to the index case. A score of 1.0 is assigned to the index case, with an additional 0.5 for each affected 1^st degree relative, and an additional 0.25 for each affected 2nd degree relative. Where an individual has bilateral cancer their score is doubled. In the GEN1 mutation screen we utilised 192 samples with a median Family History Score of 2.75 (range was 1.75–4). All cases were negative for BRCA1 and BRCA2 mutations and large deletions/duplications. In the GEN1 association study we utilised 3,750 cases with a median Family History Score of 1.75 (range 1–5.25). BRCA1/2 mutations had either been excluded (3,304) or the status was unknown (446).

Controls were obtained from the 1958 Birth Cohort Collection, an ongoing follow-up of persons born in Great Britain in 1 week in 1958 [24]. Informed consent has been obtained for all blood samples in this collection to be used as a genetic resource. Additional controls for the genomewide association study were obtained from the United Kingdom Blood Services Collection of Common Controls established for the Wellcome Trust Case Control study, a collection of DNA samples from consenting blood donors of the English, Scottish and Welsh Blood Services [25]. Individuals of self-reported white ethnicity and representative of gender and each geographical region were selected.

GEN1 mutation analyses

We screened genomic DNA from the familial breast cancer case and control samples through the 13 coding exons and intron–exon boundaries of GEN1 (Q17RS7) in 15 PCR fragments (Supplementary Table 1). Following PCR, we carried out uni-directional sequencing using BigDyeTerminator Cycle sequencing kit and 3730XL automated sequencer (ABI). All variants and mutations were confirmed by separate bi-directional sequencing in a different aliquot of native DNA. We analysed the coding sequence and ten intronic flanking bases of each exon using Mutation Surveyor software version 3.20 (SoftGenetics) and visual inspection. Only samples successfully analysed through at least 90% of the GEN1 coding sequence were included; we successfully mutationally screened 176/192 cases and 159/184 controls. We assessed the likely pathogenicity of variants using Polyphen (http://genetics.bwh.harvard.edu/pph/), SIFT (http://blocks.fhcrc.org/sift/SIFT.html) and NNSplice (http://fruitfly.org:9005/seq_tools/splice.html) in silico software.

To further evaluate the GEN1 truncating variant c.2515_519delAAGTT, we extended mutation analysis of exon 13 to 536 cases and 525 controls in total. To investigate whether the variant causes nonsense-mediated RNA decay, we extracted RNA from EBV transformed lymphoblastoid cell lines from two cases and two controls heterozygous for the variant. We used SuperScript II Reverse Transcriptase (Invitrogen) to generate cDNA which was amplified, sequenced and analysed as described above using primers Forward-AAGGAGACCAGCTGCTTCAA and Reverse-GGAAGAGGGCTATCCAAACA.

Statistical analyses

We performed comparisons of the frequencies between cases and controls of variants detected through mutational screening using a two-sided Fisher’s exact test. We carried out a genome-wide association study for breast cancer susceptibility alleles genotyping 3,960 breast cancer cases on a custom Illumina Infinium 670k array. Genotype frequencies were compared with those obtained on 5,069 controls genotyped on an Illumina Infinium 1.2M array, utilising data on 594,375 SNPs that were successfully genotyped on both arrays. We excluded closely related individuals (IBS probability >0.86), individuals with >15% non-European ancestry (by computing IBS scores between participants and individuals in HapMap and using multi-dimensional scaling) and restricted analyses to individuals that were called on >97% of successfully genotyped SNPs. After these exclusions, 3,750 cases and 4,907 controls were used in the final analysis [9].

The transcribed region of GEN1 extends from 17,798,661 to 17,830,113 bp on chromosome 2 (http://genome.ucsc.edu/) and contains 30 single nucleotide polymorphisms of minor allele frequency >0.05 (http://hapmap.ncbi.nlm.nih.gov/). Linkage disequilibrium (LD) in the region was evaluated in 90 HapMap CEU individuals using a sliding window of 1,000 kB and 10,000 SNPs. These LD data were used to select six SNPs from our dataset which tag these 30 SNPs in GEN1 at r² > 0.8 (Supplementary Table 2). We undertook association testing using a 1 df Cochran–Armitage test and a general 2 df χ² test. Analyses were performed using Stata10 (State College, TX, USA) and PLINK (v1.06) software [26].

Results

We successfully analysed the full coding sequence and intron–exon boundaries of GEN1 in 176 individuals with familial breast cancer and 159 controls (Table 1). We identified one truncating variant, c.2515_2519delAAGTT, a five base pair deletion in the final exon of the coding sequence. We extended the analysis of this mutation which demonstrated that it was present in similar frequencies in case and control chromosomes (47/1,072 cases vs. 47/1,050 controls) and both cases and controls homozygous for the deletion were identified (Fig. 1a–c). This mutation is predicted to cause protein truncation generating a product lacking 69 amino acids (~8% of the protein) from the c-terminus. The mutation is in the last exon of GEN1 and would be anticipated to escape nonsense-mediated RNA decay [27]. This was confirmed by analysis of cDNA from cases and controls heterozygous for c.2515_2519delA-AGTT, which demonstrated equal proportions of the mutant and wild-type transcripts.

Table 1.

Coding GEN1 variants in breast cancer cases and controls

Variant	dbSNPa	Allele frequenciesb		P value for associationc
		Cases	Controls
c.274T>A; p.S92T	rs1812152	195/358	226/346	0.1
c.428A>G; p.N143S	rs16981869	22/366	21/364	0.9
c.566G>A; p.S189N	–	6/362	6/328	0.9
c.607A>G; p.I203V	rs10177628	3/362	0/328	0.1
c.905G>A; p.R302H	–	2/382	1/344	0.6
c.988G>A; p.E330K	–	0/380	1/340	0.3
c.1341A>G; p.A447A	rs16983864	4/362	1/356	0.2
c.1526C>G; p.S509W	–	1/372	3/358	0.3
c.1638T>A; p.S546S	–	6/372	5/358	0.8
c.1971A>G; p.E657E	rs300168	189/384	189/350	0.5
c.2039C>T; p.T680I	rs300169	233/384	228/350	0.6
c.2445C>T; p.Y815Y	–	3/382	1/360	0.3
c.2449A>G; p.T817A	–	0/382	1/360	0.3
c.2515_2519delAAGTT	–	47/1072	47/1050	0.9
c.2567C>T; p.S856F	–	1/382	0/360	0.3
c.2619T>G; p.S873R	rs57936182	4/372	1/360	0.2
c.2644A>G; p.K882E	–	6/372	7/360	0.7
c.2692C>T; p.R898C	rs17315702	3/372	6/360	0.3

^awww.ncbi.nlm.nih.gov/projects/SNP

^bThe denominator for each variant indicates the number of chromosomes successfully sequenced

^cP value for two-sided Fisher’s exact test (1 df)

Fig. 1.

Sequence traces for wild-type deletion heterozygote and deletion homozygotes. Reverse sequencing chromatograms of the sequence encompassing the c.2515_2519delAAGTT deletion showing the wild-type sequence (a) deletion heterozygote sequence (b) and deletion homozygote sequence (c). The five deleted bases are indicated by the red square in wild-type sequence

We also identified four synonymous and 13 non-synonymous GEN1 variants. 13 variants were detected at similar frequencies in cases and controls including five common variants (frequency >0.05). Two rare non-synonymous variants were found in cases but not controls and two rare non-synonymous variants were found in controls but not cases. None of the variants were predicted to affect splicing. Only one variant, c.2692C>T p.R898C, was predicted to be deleterious by both Polyphen and SIFT algorithms but the difference in frequency between case (3/372) and control (6/360) chromosomes was not significant (P = 0.3) (Table 1).

We compared the frequency between 3,750 familial breast cancer cases and 4,907 controls of six SNPs which tag the 30 common variants in the genomic region encompassing GEN1 (Supplementary Table 2). There was no evidence of significant association for any of these tag SNPs with breast cancer (Table 2).

Table 2.

Association with breast cancer of six SNPs tagging common GEN1 variants

Illumina tag SNP	Minor allele	MAF cases	MAF controls	P valuea
rs7556886	T	0.19	0.19	0.97
rs6761104	A	0.10	0.10	0.49
rs300168	A	0.46	0.47	0.69
rs300169	G	0.36	0.36	0.42
rs17315736	A	0.09	0.09	0.49
rs13031876	C	0.35	0.35	0.60

MAF minor allele frequency

^aCochran–Armitage trend test (1 df), unadjusted for multiple testing

Discussion

GEN1 was recently identified as a Holliday junction resolvase with a key role in repair of DNA double-strand breaks. This function, together with the report of somatic GEN1 mutations in two breast cancers, led to the proposition that GEN1 would act as a breast cancer susceptibility gene, similar to some other DNA repair genes [1-5, 7]. In these recognised breast cancer susceptibility genes, BRCA1, BRCA2, CHEK2, ATM, BRIP1 and PALB2, inactivating, primarily truncating, mutations confer high or intermediate risks of breast cancer. We identified a single GEN1 truncating mutation, c.2515_2519delAAGTT. However, this deletion was present at equal frequency in cases and controls, indicating that it is not associated with appreciable increased risk of breast cancer. The deletion is in the final exon of the gene, results in truncation of <10% of the protein, and mutant transcripts are not subjected to nonsense-mediated decay. Moreover, we identified several control individuals homozygous for the deletion, demonstrating that the last 69 amino acids of the GEN1 protein are dispensable for its function. This is consistent first with findings of Ip et al. [21] who reported that a truncated form of GEN1, lacking the C-terminal, is sufficient for Holliday junction resolvase activity and secondly with phylogenetic evidence which demonstrates strong conservation between GEN1 and its yeast homologue yen1 over the first 480 amino acids, but very little in the C-terminal regions [21]. Our mutation screen did not identify any additional truncating mutations, and there was no evidence that non-truncating variants are likely to be pathogenic.

Within recent years, common variants conferring small risks of breast cancer have been identified using large case–control series via genome-wide analyses of single nucleotide polymorphisms [8-15, 28]. Of the 18 common, low-penetrance breast cancer susceptibility alleles identified to date, none have been in regions containing DNA repair genes. We evaluated 30 common SNPs in the vicinity of GEN1 by comparing the frequencies of six tag SNPs in 3,750 breast cancer cases and 4,907 controls and found no evidence to suggest that any common variants in this region are associated with breast cancer.

Our mutational screening data indicate that GEN1 does not make an appreciable contribution to breast cancer predisposition by acting as a high-penetrance breast cancer predisposition gene akin to BRCA1 and BRCA2 or intermediate-penetrance breast cancer predisposition gene, similar to ATM, BRIP1, CHEK2, or PALB2. The association analysis finds no evidence that common variation targeting GEN1 confers susceptibility to breast cancer. Overall, these data strongly suggest that constitutional GEN1 variation does not contribute to breast cancer predisposition. In addition, our analyses demonstrate the importance of undertaking appropriate genetic investigations, typically full gene screening in cases and controls together with large-scale case–control association analyses, to evaluate the contribution of genes to cancer susceptibility.

Appendix

The authors gratefully acknowledge the clinicians and counsellors from the Breast Cancer Susceptibility Collaboration UK (BCSC) who coordinated recruitment and collection of the FBCS samples: A. Ardern-Jones, G. Attard, K. Bailey, C. Bardsley, J. Barwell, L. Baxter, R. Belk, J. Berg, N. Bradshaw, A. Brady, S. Brant, C. Brewer, G. Brice, G. Bromilow, C. Brooks, A. Bruce, B. Bulman, L. Burgess, J. Campbell, B. Castle, R. Cetnarskyj, C. Chapman, C. Chu, N. Coates, A. Collins, J. Cook, S. Coulson, G. Crawford, D. Cruger, C. Cummings, R. Davidson, L. Day, L. de Silva, B. Dell, C. Dolling, A. Donaldson, A. Donaldson, H. Dorkins, F. Douglas, S. Downing, S. Drummond, J. Dunlop, S. Durrell, D. Eccles, C. Eddy, M. Edwards, E. Edwards, J. Edwardson, R. Eeles, F. Elmslie, G. Evans, B. Gibbens, C. Giblin, S. Gibson, S. Goff, S. Goodman, D. Goudie, L. Greenhalgh, J. Greer, H. Gregory, R. Hardy, C. Hartigan, T. Heaton, C. Higgins, S. Hodgson, T. Homfray, D. Horrigan, C. Houghton, L. Hughes, V. Hunt, L. Irvine, L. Izatt, L. Jackson, C. Jacobs, S. James, M. James, L. Jeffers, I. Jobson, W. Jones, S. Kenwrick, C. Kightley, C. Kirk, L. Kirk, E. Kivuva, A. Kumar, F. Lalloo, N. Lambord, C. Langman, P. Leonard, S. Levene, S. Locker, P. Logan, M. Longmuir, A. Lucassen, V. Lyus, A. Magee, S. Mansour, D. McBride, E. McCann, V. McConnell, M. McEntagart, K. McDermot, L. McLeish, D. McLeod, L. Mercer, C. Mercer, Z. Miedzybrodzka, J. Miller, P. Morrison, J. Myring, J. Paterson, P. Pearson, G. Pichert, K. Platt, M. Porteous, C. Pottinger, S. Price, L. Protheroe, L. Protheroe, S. Pugh, C. Riddick, V. Roffey-Johnson, M. Rogers, S. Rose, S. Rowe, A. Schofield, G. Scott, J. Scott, A. Searle, S. Shanley, S. Sharif, J. Shaw, J. Shea-Simonds, L. Side, J. Sillibourne, K. Simon, S. Simpson, S. Slater, K. Smith, L. Snadden, J. Soloway, Y. Stait, B. Stayner, M. Steel, C. Steel, H. Stewart, D. Stirling, M. Thomas, S. Thomas, S. Tomkins, H. Turner, E. Tyler, E. Wakeling, F. Waldrup, L. Walker, L. Walker, C. Watt, S. Watts, A. Webber, C. Whyte, J. Wiggins, E. Williams, L. Winchester.