• CHEK2*1100delC
• CHEK2
• breast cancer
• cancer risk
• genetic epidemiology
• cancer: colon
• genetic screening/counselling
• Diagnostics tests
• genetics
• cancer: breast
• epidemiology
• molecular genetics
• clinical genetics

The CHEK2 kinase functions in the DNA damage response pathway and, activated by ATM, phosphorylates a variety of targets such as CDC25A/C, BRCA1 and p53.1 Both truncating and some missense mutations in CHEK2 confer an increased breast cancer risk. Besides rare CHEK2 germline mutations, five founder mutations in CHEK2 have been identified with a variable prevalence in different populations.2 The p.S428F mutation has a prevalence of 1.37% in the Ashkenazi Jewish population, p.I157T is most prevalent in Slavic populations (∼5.0%), IVS+2G>A has been found in German and Polish populations (∼0.3%), the del5395 in the Polish population, and CHEK2*1100delC with highest prevalence in northern Europe.2 The CHEK2*1100delC mutation results in a frameshift leading to a premature termination at codon 381. In the Netherlands the prevalence in the general population is 1.1%, 2.5% in unselected breast cancer cases, and up to 4.9% in familial breast cancer cases.3 4 The population based breast cancer risk for a female CHEK2*1100delC carrier is consistently associated with odds ratios between 1.5 and 3.0, corresponding to an estimated lifetime risk for breast cancer of approximately 20–25%.2–5

Although arguably influenced by ascertainment biases, publications have shown that female heterozygous CHEK2*1100delC carriers from a familial breast cancer setting have a higher lifetime breast cancer risk of 37%.6 7 Since females from western populations have an average lifetime risk of ∼10%, this suggests that these females have a threefold increased breast cancer risk. This increased risk in a familial setting is most likely explained by a polygenic model in which other genetic risk factors present in these families contribute to the individual risk of these heterozygous cases.6 Clinical utility of CHEK2*1100delC mutation testing is currently under debate.2 8 9

We considered the question of whether homozygous female CHEK2*1100delC carriers have a higher cancer risk than heterozygous carriers. Since the prevalence of the CHEK2*1100delC allele is high in the Netherlands we are in the unique position to address this question. The prevalence of homozygous individuals in the general population is estimated to be approximately 1 in 33 000 individuals.

Genomic DNA of 2554 independent breast cancer patients from familial non-BRCA1/2 breast cancer cohorts of three Dutch Clinical Genetics Centres were genotyped for the CHEK2*1100delC mutation. For the majority of families, DNA from only one affected individual was available. If DNA was available from multiple breast cancer patients of a family, the case with the earliest onset of breast cancer was selected. The frequency of the CHEK2*1100delC allele in the general population was estimated based on the genotypes of 3267 controls. No homozygous and 37 heterozygous individuals were detected among controls. Of these controls, 2059 were derived from two previous Dutch studies and the remaining 1208 controls were genotyped for this study.10 11 Material of patients was used following institutional guidelines; breast cancer cases gave informed consent to screen for susceptibility genes. DNA from irreversibly anonymised controls was used following the code for Proper Secondary Use of Human Tissue.

Heterozygosity for the CHEK2*1100delC mutation was identified in 112/2554 female breast cancer cases (4.4%) and we found 8/2554 cases to be homozygous for the allele (0.3%). Homozygosity for CHEK2*1100delC was confirmed by sequencing. We estimated the genotype relative risk (GRR) based on the observed allele frequencies in our cases and controls and under the assumption of the Hardy-Weinberg equilibrium (HWE) in the total population. The estimated GRR for heterozygous and homozygous familial breast cancer cases were 4.04 (95% CI 0.88 to 21.0) and 101.34 (95% CI 4.47 to 121 000), respectively. The GRR for homozygotes is much higher than that of heterozygotes but the confidence intervals are broad and overlapping. We therefore tested and confirmed the hypothesis that the breast cancer risk of homozygous females was more than twice the risk of heterozygous CHEK2*1100delC females (p=0.044), showing that two CHEK2*1100delC alleles result in a more than additive risk. For detailed information about the statistical methods and analysis see online appendix. If we interpret the results in a conservative way, it indicates that the breast cancer risk for homozygous CHEK2*1100delC female carriers is increased more than fourfoldthat is, more than twice the risk of heterozygous carriers in the population. Since the risk estimates are based on the comparison of risks within a familial cohort, the risk of familial homozygous cases is likely to be more than sixfold—that is, more than twice the risk of heterozygous familial cases. However, even the conservative estimate justifies intensive breast cancer surveillance.

We then tested all available siblings (blood DNA or DNA from paraffin embedded material when deceased) of the eight homozygous cases for the presence of CHEK2*1100delC homozygosity (for pedigrees see figure 1). Of the total of 26 individuals tested in the generation concerned, we identified a total of 12 homozygous individuals (two males, 10 females, including our initial eight probands), and 12 heterozygous individuals (four males, eight females). All 10 female homozygotes had breast cancer; 7/10 had multiple primary tumours, four had bilateral breast cancer, and four were additionally affected with colon cancer, ovarian cancer, uterine cancer or melanoma, respectively (table 1). Of note, no homozygosity was found in any of the six unaffected females. One homozygous male had a thymoma at age 47 years and the other homozygous male was cancer-free at age 54 years.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Pedigrees of CHEK2*1100delC homozygous breast cancer patients. The initial eight probands with CHEK2*1100delC homozygosity (+/+) are indicated by arrows. Genotyping results are mentioned (+/− =heterozygosity and −/− =wild-type) above each individual. Individuals with breast cancer (B) are shown as filled circles, with the age at diagnosis. Other cancers are indicated beneath the relevant individuals. Carc, carcinoma of unknown type; Co, colon; Cx, cervical; Eso, esophageal; Leu, leukaemia; Lu, lung; Mel, melanoma (Oc, ocular; is, in situ); NA, not available; Ov, ovarian, Pa, pancreatic; Re, renal; Sarc, chondrosarcoma; Sk, skin; Thym, thymoma; Ut, uterine; and d, age of death. Family 4 was lost to follow-up. Informed consent was obtained from all homozygous females and their family members following institutional guidelines. Additional test results are indicated below the individual involved.

In the proband generation 12 heterozygous CHEK2*1100delC carriers were identified, of whom four were affected with cancer. Of the 36 siblings in total 10 (five males and five females) were not available for testing (six from one family, family 4).

Table 1 shows the age of onset and histology of the breast cancer tumours in homozygous females. Median age of diagnosis of (first) breast cancer for the homozygous females was 47 years (average age 44 years).

Of the 12 available tumours, 11 were grade 2–3 invasive ductal carcinoma. Nine out of 12 (75%) breast cancers were oestrogen receptor positive and eight progesterone receptor positive tumours, confirming the predominantly hormone receptor positive tumours observed in heterozygote carriers of the CHEK2*1100delC mutation (table 1).12 Remarkably, HER2 staining was positive in six out of 12 tumours (50%) in comparison to the known 20–25% in breast cancer tumours in the population.

Figure 1 shows the pedigrees of the CHEK2*1100delC homozygous carriers. Additional DNA testing, besides BRCA1/2 screening, was performed in some families to exclude other known tumour syndromes (for test results see pedigrees). No other DNA diagnosis could be made except for the homozygosity for the CHEK2*1100delC allele in these families.

In previous studies, homozygosity for the CHEK2*1100delC allele was once described in a Dutch male with colorectal cancer at age 52 years and in a patient from German/Welsh descent who was bilaterally affected with breast cancer at age 47 and 61 and had an uterine sarcoma at age 58 years.5 13 In an independent bilateral breast cancer study we enrolled one additional family with homozygosity of the CHEK2*1100delC mutation and identified one homozygous female with bilateral breast cancer age 50 and 56 (both ER/PR positive, HER2 negative), and colon cancer at age 59; and two homozygous males of whom one died of colon cancer at age 32 years. These cases match well with our findings reported here. It is remarkable that colon cancer thus far has been described (present study combined with literature and our additional family) in four of the 17 homozygous individuals, with ages of onset of 32, 43, 59, and 52 years. Publications variably showed a higher prevalence of heterozygosity for the CHEK2*1100delC mutation in colorectal cohorts.14 15 Systematic studies are needed to evaluate the association between colorectal cancer or other cancers and CHEK2*1100delC homozygosity.

This is the first paper to describe the incidence of homozygosity for the CHEK2*1100delC in a large familial non-BRCA1/2 breast cancer cohort.

In summary, CHEK2*1100delC homozygous females in a familial breast cancer setting have a high breast cancer risk. A conservative estimate shows that the risk is more than fourfold higher than the risk for women in the general population. However, since the breast cancer risks are based on the GRR estimates of heterozygous and homozygous women within a familial breast cancer cohort, the risk for homozygous women in this setting is likely more than sixfold increased. The GRR estimate of 101.34 (95% CI 4.47 to 121 000) for homozygotes also supports a breast cancer risk substantially higher than four. It is reasonable to assume that similar risks will apply for women that are homozygous or compound heterozygous for other deleterious mutations in CHEK2, such as IVS+2G>A and del5395. Further studies are needed to elucidate if this also applies for missense variants such as p.I157T and p.S428F. Interestingly, the phenotype of homozygous CHEK2*1100delC individuals is different from biallelic mutation carriers in the other known moderate breast cancer risk genes which cause specific syndromes. Recessive inheritance of mutations in PALB2 and BRIP1 result in Fanconi anaemia and in ATM results in ataxia telangiectasia.

We conclude that diagnostic testing for mutations in CHEK2 has clinical utility in a familial breast cancer setting, especially in populations where these risk alleles are relatively prevalent. Based on our study, we propose that women who are homozygous or compound heterozygous for deleterious mutations in CHEK2 are eligible for intensive breast surveillance in a specialist environment concordant with guidelines for female BRCA1/2 mutation carriers. If our data are replicated and a level of risk similar to BRCA1/2 carriers is further validated in homozygous CHEK2 carriers, these women should also have access to preventive mastectomy after adequate counselling.