Molecular profile of ductal carcinoma in situ of the breast in BRCA1 and BRCA2 germline mutation carriers
- 1Department of Pathology, University Medical Center Utrecht, Utrecht, The Netherlands
- 2Department of Internal Medicine, University Medical Center Utrecht, Utrecht, The Netherlands
- 3Department of Clinical Genetics, VU University Medical Center, Amsterdam, The Netherlands
- 4Department of Pathology, University Medical Center Groningen, Groningen, The Netherlands
- 5Department of Medical Oncology, University Medical Center Groningen, Groningen, The Netherlands
- Correspondence to Petra van der Groep, Department of Pathology, University Medical Center Utrecht, PO Box 85500, 3508 GA Utrecht, The Netherlands;
- Accepted 2 June 2009
- Published Online First 18 June 2009
Aims: Ductal carcinoma in situ (DCIS) is an established late precursor of sporadic invasive breast cancer and to a large extent parallels its invasive counterpart with respect to molecular changes and immunophenotype. Invasive breast cancers in germline BRCA1 and BRCA2 mutation carriers have a distinct “basal” and “luminal” immunophenotype, respectively, but the immunophenotype of their precursor lesions has hardly been studied, and this was the aim of this study.
Methods: DCIS lesions of 25 proven BRCA1 and 9 proven BRCA2 germline mutation carriers and their 22 and 6, respectively, accompanying invasive lesions were stained by immunohistochemistry for oestrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor (HER)2/neu, cytokeratin (CK)5/6, CK14, epidermal growth factor receptor (EGFR) and Ki67.
Results: DCIS lesions in BRCA1 mutation carriers were mostly of the basal molecular type with low ER/PR/HER2 expression, while they frequently expressed CK5/6, CK14 and EGFR, and were mostly grade 3 and highly proliferative. DCIS lesions in BRCA2 mutation carriers were mostly of luminal molecular type with frequent expression of ER/PR, and infrequent expression of CK5/6, CK14 and EGFR, and they were mostly grade 3 and showed low proliferation. In BRCA1 and BRCA2 mutation carriers there was a high concordance between DCIS lesions and their concomitant invasive counterpart with regard to expression of individual markers as well as “molecular” subtype.
Conclusions: Although the number of cases studied was low, DCIS lesions in BRCA1 and BRCA2 mutations carriers are usually of the basal and luminal molecular type, respectively, similar to their accompanying invasive cancers, thereby providing evidence that DCIS is a direct precursor lesion in these hereditary predisposed patients. This also suggests that crucial carcinogenetic events leading to these phenotypes in hereditary predisposed patients occur before the stage of invasion.
Carriers of germline mutations in BRCA1 or BRCA2 have a hereditary predisposition for developing breast1 2 and/or ovarian/Fallopian tube cancer.3 4 Several studies have indicated that the genetic makeup of BRCA1/2-related breast cancers is different from that of sporadic breast cancer. These differences comprise gains and losses of specific parts of chromosomes as well as differences in gene expression.5 6 7 8 9 10 11 In line with this, the morphological and immunohistochemical phenotype of BRCA1-related and BRCA2-related breast cancers is also different.12 13 14
The majority of the BRCA1-related cancers are high-grade poorly differentiated ductal carcinomas,15 16 although a proportion of these cancers show a medullary differentiation. Most BRCA1-related cancers are well demarcated and show a conspicuous lymphocytoplasmic infiltrate. The immunophenotype of these cancers comprises low expression of oestrogen receptor (ER), progesterone receptor (PR) and human epidermal growth factor receptor (HER)2/neu receptors.17 They often lack p27Kip1 and cyclin D1 expression,13 18 19 but frequently accumulate p5320 and overexpress cyclin E,18 cytokeratin (CK)5/620 21 and CK14,22 23 epidermal growth factor receptor (EGFR),24 25 26 hypoxia inducible factor (HIF)-1α27 and p-cadherin,28 in general referred to as “of basal type”.
BRCA2-related breast cancers are usually of ductal type, although a higher incidence of lobular tubular and cribriform cancers has been described compared with BRCA1-related cancers. BRCA2-related cancers are moderately to poorly differentiated carcinomas.29 30 31 32 The immunophenotype of BRCA2-related breast cancers resembles more that of sporadic cancers with frequent expression of ER, PR and only “luminal” CKs, such as CK8/18.33
Sporadic invasive breast cancer is generally thought to derive through different morphologically recognisable pre-invasive stages such as atypical hyperplasia and ductal carcinoma in situ (DCIS), and the molecular profile and immunophenotype of DCIS parallels that of its invasive counterpart.34 35 36 DCIS is relatively rarely found next to BRCA1-related and BRCA2-related invasive breast cancer, but DCIS, fibroadenoma and ductal hyperplasia seem to be more common in prophylactic mastectomy (PM) specimens of BRCA1 and BRCA2 mutation carriers than observed in control mammoplasty specimens.16 29 37 38 39 40 41 42 43 44 45 Lymphocytoplasmic infiltrate, often seen in hereditary invasive cancers, has also been observed in DCIS of hereditary patients,46 and T cell lobulitis is seen frequently in PM specimens of these women.47 No studies have so far been performed on the immunophenotype or the molecular profile of pre-malignant breast lesions in BRCA1 and BRCA2 germline mutation carriers. The aim of this study was therefore to examine the immunophenotype of DCIS in BRCA1 and BRCA2 mutation carriers and to compare this profile with its invasive counterpart (if available) to establish DCIS as a precursor of hereditary breast cancer.
Materials and methods
The original study group comprised 134 patients with invasive breast cancer and a known BRCA1 or BRCA2 mutation. Of these, DCIS material was available from 25 and 9 patients with BRCA1 and BRCA2 germline mutation, respectively. For some women no invasive cancer tissue was left in the available blocks; this left a study group comprising 22 BRCA1 and 6 BRCA2 patients with concomitant invasive cancer. Paraffin-embedded blocks of these patients were derived from the archives at the University Medical Center Utrecht, the VU University Medical Center, Amsterdam, and the University Medical Center Groningen, The Netherlands. Use of anonymous or coded leftover material for scientific purposes is part of the standard treatment contract with patients in The Netherlands.48
Tumour samples had been fixed in neutral buffered formaldehyde, and processed to paraffin blocks according to standard procedures. Sections (4 μm thick) were cut and stained with H&E for histopathology. DCIS grading was performed by consensus of two experienced breast pathologists, according to Holland et al.49
After deparaffinisation and rehydration, antigen retrieval was performed at boiling temperature in citrate buffer pH 6 (ER, PR, HER-2/neu, Ki67) or EDTA buffer pH 9 (CK5/6, CK14). For EGFR, antigen retrieval was performed by incubating the slides with Prot K solution (ready to use; Dako, Glostrup, Denmark) for 5 min at room temperature. A cooling off period of 30 min preceded the incubation (60 min at room temperature) with the following mouse monoclonal antibodies: ER (1:50; Dako), PR (1:50; Novocastra, Newcastle upon Tyne, UK), HER-2/neu (1:100; Neomarkers, Lab Vision Corporation, Fremont, California, USA), CK5/6 (1:3000; Chemicon, Temecula, California, USA), CK14 (1:400; Neomarkers), proliferation marker Ki67 (1:40, MIB-1; Immunotech, Marseille Cedex, France). For EGFR (1:30; Zymed, South San Francisco, California, USA) the incubation was done overnight at 4°C. For detection of the primary antibodies a goat anti Ms/Rb/Rt-poly-HRP (ready to use; Powervision, Immunologic, Immunovision Technologies, Brisbane, California, USA) was used, except for EGFR for which a Novolink Max Polymer detection system (ready to use; Novocastra) was applied. All slides were developed with diaminobenzidine followed by haematoxylin counterstaining. Before the slides were mounted, all sections were dehydrated in alcohol and xylene.
Scoring was performed by one experienced observer (PJvD) blinded to the origin of the breast lesion on the available DCIS ducts (usually >5) and one section of invasive cancer. For ER, PR and Ki67 the percentage of positively stained nuclei was estimated semiquantitatively. Cases with 10% or more nuclei stained were denoted ER/PR positive, and DCIS cases with ⩾25% and invasive cancers ⩾35% Ki67 staining were denoted as high proliferation (based on median values). HER2/neu was scored according to the DAKO system, regarding 3+ cases as positive. EGFR was scored positive when a clear membrane staining pattern was seen, and CK5/6 and CK14 when clear cytoplasmic staining was observed.
DCIS and invasive lesions were also classified as luminal (ER and/or PR positive), HER2 (HER2 positive and negative for ER and PR) or basal (CK5/6 or CK14 or EGFR positive and negative for ER, PR and HER2) according to the immunohistochemical surrogate of the new “molecular” classification of breast cancer originally based on gene expression profiling21 50 51 52 53
As shown in table 1, DCIS in BRCA1 cases (n = 25) showed very frequent expression of CK5/6 (72%), frequent expression of CK14 (44%) and EGFR (44%), while expression of ER (32%), PR (16%) and HER2 (0%) was infrequent; 21/25 (84%) cases were CK5/6 or CK14 or EGFR positive. The mean percentage of Ki67 staining in these cases was 27% (range 0–100). Nine cases were grade 2, and 16 were grade 3. As to the molecular classification, eight BRCA1-mutated DCIS cases were of the luminal type and 17 cases were categorised as basal. When comparing BRCA1-related DCIS with its invasive counterpart (n = 22), expression of ER and PR was similar in all cases (table 2), while HER2, CK5/6, CK14, EGFR, and Ki67 showed concordance in 91%, 91%, 77%, 77% and 55% of cases, respectively. As to the molecular classification, 7 and 14 cases were luminal and basal type, respectively, concordant in DCIS and invasive cancers, while one case was classified as HER2 in the invasive lesion and basal in DCIS (concordance 95%) (see table 3).
DCIS in BRCA2 cases (n = 9) showed very frequent expression of ER (89%), frequent expression of PR (44%), while expression of HER2 (33%), EGFR (33%), CK5/6 (11%) and CK14 (11%) was infrequent (table 1). Three out of nine cases (33%) were CK5/6 or CK14 or EGFR positive. The mean percentage of Ki67 staining in these cases was 14% (range 0–100). Three cases were grade 2, and six were grade 3. As to the molecular classification, all the cases were luminal. When comparing BRCA2-related DCIS with its invasive counterpart (n = 6), expression of Ki67 was similar in all cases, while the concordance of expression of HER2, CK5/6, CK14, ER, PR and EGFR was 83%, 83%, 83%, 66%, 66% and 50%, respectively (table 4). As to the molecular classification, five cases were luminal concordant in the DCIS and invasive lesion, and one case was HER2 in the invasive lesion and luminal in the DCIS (concordance 83%) (table 3).
The molecular subtype and immunophenotype of invasive sporadic and BRCA1-related and BRCA2-related breast cancers has been fairly well unravelled. However, direct breast cancer precursor lesions and genetic progression routes have only been identified for sporadic breast cancer, with DCIS as the final stage before invasion takes place.34 35 36 37 The aim of the present study was to evaluate the immunophenotype of DCIS lesions in BRCA1 and BRCA2 mutation carriers and to compare this profile with their invasive counterparts (if available) to establish whether DCIS is a precursor of hereditary breast cancer.
We were able to compile a group of 25 and 9 DCIS lesions, respectively, in BRCA1 and BRCA2 germline mutation carriers from an original group of 134 patients with BRCA1-related and BRCA2-related invasive breast cancer derived from three academic centres. The percentage of cases DCIS in this group is thereby relatively low (20%) compared with studies in non-BRCA mutation carriers (40% or even 59%),16 45 indicating that DCIS is rarely present next to invasive BRCA1-related and BRCA2-related cancers. These small numbers of cases obviously do not allow definite conclusions.
In the present study, DCIS lesions in BRCA1 mutation carriers were mostly of the basal type with low expression of ER, PR and HER2, while frequently expressing CK5/6, CK14 and EGFR. Further, they were mostly grade 3 and showed high proliferation. This is a similar immunophenotype to that described for invasive BRCA1-related invasive breast cancers.13 25 26 Comparing the DCIS immunophenotype with the invasive lesion of the same patient, a high similarity in immunophenotype and molecular subtype was observed. DCIS lesions in BRCA2 mutation carriers were mostly of luminal type with frequent expression of ER and PR, and infrequent expression of CK5/6, CK14 and EGFR. Further, they were mostly grade 3 and showed low proliferation. This is also similar to the immunophenotype of invasive BRCA2-related breast cancers earlier described.33 The present study showed a high concordance between BRCA2-related DCIS lesions and their concomitant invasive counterparts with regard to expression of individual markers as well as molecular subtype. This illustrates that DCIS is very likely to be a direct precursor lesion of invasive BRCA1-related and BRCA2-related breast cancers, with BRCA1 mutation carriers mostly following the basal progression route, while BRCA2 mutation carriers predominantly follow the luminal progression route. The discrepancies in immunophenotype and molecular phenotype between DCIS lesions and invasive counterparts may largely be explained by the inherent heterogeneity of neoplasias. However, in view of the relatively low number of BRCA2-related DCIS lesions that were included in the present study, these results have to be interpreted with more care.
These observations suggest that crucial carcinogenetic events leading to these apparently dedicated phenotypes occur in BRCA1 and BRCA2 mutation carriers before the stage of invasion. However little is known about the earliest precursor lesions for BRCA1-realted and BRCA2-related breast cancers, although a variety of lesions proposed to be precursor lesions for sporadic breast cancer have been described in increased frequency in PM specimens in patients with a hereditary predisposition. These lesions encompass fibroadenoma, and atypical ductal and lobular hyperplasia.37 38 39 40 41 42 43 44 In view of the luminal nature of BRCA2-related more advanced lesions, it may well be that these are preceded by luminal type precursors such as cylindrical cell lesions, atypical ductal hyperplasia and lobular neoplasia. These lesions are, on the other hand, unlikely precursors in the basal BRCA1-related progression route. The search for earlier basal precursors is therefore on. Whether it is that a subgroup of ductal hyperplastic lesions may have clonal basal properties and thereby be a precursor of BRCA1-related DCIS, or that BRCA-related DCIS almost directly derives from breast stem cells, is yet a matter of speculation.
Since the BRCA1 or BRCA2 phenotype is apparently already present in the pre-invasive stage, preventive systemic treatment should take this into account. While tamoxifen may help to prevent or delay development of DCIS/invasive cancer in patients with BRCA2 mutation, it may be ineffective in BRCA1 carriers that may be more sensitive to EGFR inhibitors. Further, targeted imaging strategies to detect DCIS lesions in these hereditary patients in an early phase by, for example, positron emission tomography or optical molecular imaging techniques using targeted tracers may have to be fine tuned based on molecular phenotype.
Ductal carcinoma in situ (DCIS) lesions of BRCA1 mutation carriers are usually of basal molecular type.
DCIS lesions of BRCA2 mutation carriers are usually of luminal molecular type.
DCIS lesion of BRCA1 and BRCRA2 mutation carriers are similar to accompanying invasive lesions.
Crucial carcinogenetic events leading to the characteristic phenotype of breast lesions in BRCA1 and BRCA2 mutation carriers occur before the stage of invasion.
In conclusion, the immunophenotype and (consequently) the molecular phenotype of DCIS in BRCA1 and BRCA2 mutation carriers are similar to their accompanying invasive cancers, thereby providing evidence that DCIS is a direct precursor lesion in these hereditary predisposed patients. DCIS lesions in BRCA1 mutation carriers are usually of the basal type, while DCIS in BRCA2 mutation carriers is predominantly of the luminal phenotype. This suggests that crucial carcinogenetic events leading to these phenotypes occur before the stage of invasion. Further studies are required to confirm these results in larger groups, and to identify the earliest precursors of BRCA1 and BRCA2-related breast cancer.
We thank Dr Carolien van Deurzen for help with grading of the DCIS cases.
Funding Partly supported by an unrestricted educational grant from Aegon, Inc.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.