Along with environment and lifestyle, a woman’s genetic background contributes to her risk of having breast cancer. Her risk of developing breast cancer increases if she has a family history of the disease. In a re-analysis of 52 studies comparing cancer incidence in relatives of breast cancer cases and in controls, RR for breast cancer increased with increasing numbers of affected first-degree relatives: compared with women who had no affected relative, the RR was 1.80, 2.93, and 3.90, respectively, for women with one, two, and three or more affected first-degree relatives [
65]. RR was greatest at young ages and, for women of a given age, was greater the younger the relative was when diagnosed. Twin studies provide evidence that genes contribute substantially to this excess familial risk of breast cancer. Combined data on all twin pairs listed in three North-European registries showed that concordance for breast cancer was two times higher among monozygotic twins, who share all their genes, than among dizygotic twins, who share half their genes [
66].
Four high-penetrance genes are tested in clinical practice when genetic susceptibility to breast cancer is suspected:
BRCA1,
BRCA2,
TP53 and
CDH1. Germline mutations in
BRCA1 and
TP53 are predominantly associated with IDC,
BRCA2 mutations are associated with both ductal and lobular tumors, while mutations in
CDH1 are exclusively associated with ILC. Mutations in
PTEN and
STK11 cause, respectively, Cowden and Peutz-Jaeger syndrome, and breast cancer risk is also high in affected females. However, the presenting manifestations are usually not related to breast cancer (for example, macrocephaly, central nervous system abnormalities, mucocutaneous lesions, gastrointestinal hamartomas) and we shall therefore not discuss these two syndromes further [
69,
70].
BRCA1, BRCA2 and TP53
In the 1990s, linkage analyses and positional cloning in breast and breast-ovarian cancer families led to the identification of
BRCA1 and
BRCA2 [
71,
72]. Both have a role in maintaining DNA integrity. Mutations confer a high risk of breast and ovarian cancer with estimated breast cancer penetrances of 60% for
BRCA1 and 55% for
BRCA2 by age 70 years [
73]. However, not all breast and breast-ovarian cancer families carry a mutation in
BRCA1 or
BRCA2 [
74]. For example, 41% of families with four or five cases of breast cancer under the age of 60 years - but no ovarian cancer - are carriers, but that proportion increases to 88% in families with at least four breast cancer cases and one case of ovarian cancer. Mutations are rare in population-based, unselected breast cancer cases. The detection rate was, for example, 2% in a large English series of women diagnosed below age 55 years, although it increased to 12% in the subset of women diagnosed before age 35 years [
75]. These are likely underestimates, as the sensitivity of gene analysis techniques was lower in the late 1990s than it is today. Some phenotypic characteristics influence the probability of carrying a mutation in
BRCA1 or
BRCA2. For example, up to 15% of unselected women with triple-negative breast cancer have a
BRCA1 mutation, while there does not seem to be an association with
BRCA2 [
76]. Similarly, the distribution among different breast cancer types varies according to the predisposing gene involved. The CIMBA Consortium analyzed the pathology of invasive breast cancers in 6,893
BRCA1/
2 mutation carriers, and found that only 2.2% of tumors associated with
BRCA1 were ILC. In contrast, the proportion of ILCs in
BRCA2 mutation carriers was 8.4%, closer to the characteristics of breast cancers from the general population [
1,
77].
Li-Fraumeni syndrome is characterized by early onset of a variety of tumors. It is caused by mutations in the tumour-suppressor gene
TP53. Affected individuals are at increased risk of sarcoma, premenopausal breast cancer, brain cancer, adrenocortical cancer, leukaemia, lymphoma, germ cell tumor, melanoma, lung cancer and cancer of the digestive tract [
78]. Cancer risk by age 45 years is about 41% in males and 84% in females; lifetime risk is 73% in males, and approaches 100% in females [
79]. The majority of cancers in females are breast cancers, and most breast cancers are diagnosed before age 45 years. Little is known regarding the histological characteristics of breast cancers associated with germline
TP53 mutations, but the two studies that have examined the issue have only shown tumors of the ductal type - and none of the lobular type - out of a total of 48 cancers in mutation carriers [
80,
81]. These data suggest that
TP53 might predispose exclusively to IDC, and not to ILC. Confirmatory studies are nevertheless needed, and it seems premature at this stage to exclude ILC from the tumor spectrum associated with Li-Fraumeni syndrome.
ILCs are therefore very much underrepresented in carriers of
BRCA1 and
TP53 mutations, while their frequency in
BRCA2 mutation carriers is more similar to that in the general population. This overall underrepresentation of lobular cancers contrasts with observations made in over 40,000 Utah cases with genealogical records showing unusually high levels of familial clustering for ILC, and therefore a higher contribution of genetic, inherited factors compared with IDC [
82]. The
CDH1 susceptibility gene likely explains at least some of this excess in familial risk, the remainder being accounted for by genes and loci that are yet to be discovered.
CDH1
CDH1 is located on chromosome 16q22 and codes for the E-cadherin protein. E-cadherin maintains tissue integrity as it mediates cell-cell adhesion. There is also evidence that forced expression of the protein inhibits the growth of breast cancer cells via mechanisms that are yet to be determined, and that the protein therefore controls cell proliferation in addition to its anti-invasion properties [
83,
84]. Its tumor-suppression role is limited to breast cancer of the lobular type. Indeed, loss of expression is observed in the majority of lobular breast carcinomas, and, in the few tumors with conserved expression, E-cadherin integrity is impaired [
85]. On the contrary, expression is unaffected in ductal breast carcinomas [
86]. First-event somatic mutations, with subsequent loss of heterozygosity or promoter methylation, are classically responsible for
CDH1 inactivation following the two-hit loss-of-function model [
84]. Therefore, an individual with an inherited, germline mutation in
CDH1 is at increased risk of ILC as a single somatic event is sufficient to generate tumorigenesis.
Readers should be reminded here that
CDH1 was initially known as a susceptibility gene for gastric cancer of the diffuse type, following the identification of germline mutations in Caucasian, Maori and African-American families with multiple affected individuals [
87-
91]. Like in ILC, E-cadherin inactivation is an early event in diffuse gastric cancer development and, as expected in this context, the histopathological characteristics of diffuse gastric cancer show similarities with ILC, with neoplastic cells permeating the mucosa and wall as scattered individual signet-ring cells or small clusters in an infiltrative growth pattern [
87,
92,
93]. In a collaborative study based on 11
CDH1 families, the International Gastric Cancer Linkage Consortium showed that the clinical penetrance for diffuse gastric cancer was high, as the estimated risk for carriers of developing the disease was 67 to 83% [
94]. Interestingly, that same study observed that, in addition to diffuse gastric cancer, female carriers were also at high risk of ILC [
94]. Indeed, there were seven cases of breast cancers in these 11
CDH1 families, some of them at an early age, and histology, when documented, was systematically of the lobular type. The estimated risk for ILC was 39% by age 80 years. Subsequent studies of families with
CDH1 mutations led to similar conclusions: in four families with a total of 22 breast cancers, all invasive tumors for which a pathological report was available were lobular [
95,
96]. As for penetrance, a recent estimate derived from 67 mutation-positive families is 56% (P Kaurah and D Huntsman, personal communication).
There is increasing evidence that a personal history of early-onset bilateral ILC or family history of multiple ILC at a young age, in the absence of diffuse gastric cancer in the family, can be associated with
CDH1 germline mutations. Masciari and colleagues [
97] described the case of a woman carrier with unilateral ILC at age 42 years, and whose mother had been diagnosed with the same condition at age 28 years. We reported three female cases who presented with bilateral ILC below age 50 years and turned out to carry mutations in
CDH1 [
98]. In the only systematic study of women with bilateral lobular breast neoplasia before age 60 years (ILC and/or lobular carcinoma
in situ), Petridis and colleagues [
99] found mutations in 4 out of 50 (8%) women. Schrader and colleagues [
100] had previously looked into the issue with discrepant findings, as they only found mutations or potentially causal variants in 4 out of 318 (1%) women with ILC either before age 45 years or regardless of age if there was a family history of breast cancer. However, it was not known how many women actually had a family history of breast cancer of the lobular type, as histology in relatives was not specified. Furthermore, there was no upper age limit for women with familial ILC, and
BRCA1/
2 mutations had not been excluded in all cases. Cancer geneticists should therefore consider prescribing
CDH1 germline analysis in patients with a personal or family history of multiple pathologically proven early-onset ILC, but no diffuse gastric cancer, as the identification of a mutation would have direct and dramatic clinical implications. The patient would be offered risk-reducing gastrectomy (assuming her ILC has been successfully treated), given the high risk of diffuse gastric cancer [
94,
101]. Her adult relatives would then undergo targeted genetic analysis to see if they carry the mutation, and those who do would also be offered risk-reducing gastrectomy. Surveillance with upper endoscopy is a poor alternative to prophylactic surgery, except in very specific situations (for example, young athletes wishing to delay surgery for professional reasons, and elderly or frail patients), as this screening modality frequently misses foci of diffuse carcinoma in mutation carriers even when accompanied by multiple random biopsies [
102,
103]. Large, multicenter studies on the prevalence of
CDH1 mutations in patients and families with multiple cases of ILC are needed.
The high risk of ILC in females carrying a
CDH1 mutation justifies personalized, intensive surveillance. The consensus 2010
CDH1 paper recommended that breast cancer surveillance be carried out within specific research protocols, and suggested annual magnetic resonance imaging (MRI) and mammogram starting at age 35 years [
101]. As ILC risk is close to the overall breast cancer risk seen in carriers of
BRCA1/
BRCA2 mutations, it seems reasonable to offer the same type of surveillance as a routine procedure, and start screening at age 30 years with annual MRI and mammogram [
104,
105]. Risk-reducing mastectomy could be an alternative. Updated international recommendations on the management of
CDH1 mutation carriers that will address the issue are expected soon.