Background: Secretory breast cancer (SBC) is a rare entity characterised by indolent clinical behaviour, distinctive histological features and the presence of a recurrent chromosomal translocation t(12;15)(p13;q25), leading to the formation of the ETV6–NTRK3 fusion gene.
Aim: To describe the molecular genetic features of a case of SBC which harbours a duplication of the t(12;15) translocation.
Methods: Tiling path array comparative genomic hybridisation (aCGH) analysis and fluorescence in situ hybridisation (FISH) using in-house-generated probes for ETV6, NTRK3 and the fusion genes, centromeric probes for chromosomes 12 and 15, and a commercially available split-apart ETV6/NTRK3 probe.
Results: FISH revealed the presence of a duplication of the translocation t(12;15), which resulted from the gain of one copy of the derivative chromosome der(15)t(12;15), retention of one normal copy of both ETV6 and NTRK3 genes and deletion of the derivative chromosome der(12)t(12;15). Consistent with FISH findings, aCGH revealed copy number gains of ETV6 and NTRK3 and deletions encompassing the regions centromeric to ETV6 and telomeric to NTRK3. Additional regions of copy number changes included gains of 10q21, 10q26.3, 12p13.3–p13.31 15q11–q25.3 and 16pq and losses of 6q24.1–q27, 12p13.2–q12 and 15q25.3–q26.3.
Conclusions: To the best of our knowledge, this is the first time a carcinoma has been shown to harbour a duplication of the ETV6–NTRK3 translocation. The presence of an additional copy of the derivative chromosome der(15)t(12;15) coupled with deletion of the other derivative der(12)t(12;15) in the modal population of cancer cells suggests that this was either an early phenomenon or conferred additional growth advantage on neoplastic cells.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Secretory breast carcinoma (SBC) is a rare low-grade malignancy, which was originally described as juvenile carcinomas in children by McDivitt & Stewart.1 However, subsequent reports suggested that only a third of the published cases occurred in children, and the median age for patients with an SBC is 25 years.1–26 SBCs account for <1% of invasive breast carcinomas and are characterised clinically by late local recurrences and prolonged survival even with lymph node involvement.5 7 10–12 Distant metastases are exceedingly rare.10 SBCs usually present as a well-circumscribed mobile mass in the subareolar region, but can occur anywhere in the breast. Histologically, SBCs are characterised by partially circumscribed nodules composed of cells arranged in solid, microcystic (honeycomb) and tubular structures. The neoplastic cells are of low nuclear grade, occasionally with vesicular chromatin and discrete nucleoli. Their cytoplasm is vacuolated, sometimes containing dense, eosinophilic secretion in the centre of the vacuole. Extracellular secretory material positive for periodic acid/Schiff (PAS) and alcian blue,5 similar to that found in the intracellular compartment, is a characteristic feature of these tumours.
Tognon et al25 showed that SBCs consistently harbour the chromosomal translocation t(12;15)(p13;q25), which results in a fusion of the ets variant gene 6 (ETV6) gene mapping to chromosome 12 and the neurotrophic tyrosine kinase receptor 3 (NTRK3) gene located on chromosome 15. The ETV6 gene, also known as TEL oncogene, encodes an E26 transformation-specific (ETS) transcription factor that is required for haematopoiesis, early angiogenesis and maintenance of the developing vascular network. ETV6 is characterised by its C-terminal domain, ETS DNA-binding domain and the N-terminal helix–loop–helix oligomerisation domain, and has been found to form oncogenic fusion genes with different gene partners including c-abl oncogene 1, receptor tyrosine kinase (ABL) gene,27 Janus kinase 2 (JAK2) gene,28 NTRK3,13 runt-related transcription factor 1 (RUNX1, aka AML1) gene,29 ecotropic viral integration site 1 (EVI1) gene,30 meningioma 1 (MN1) gene31 and caudal type homeobox 2 (CDX2) gene.32 ETV6 protein is expressed in normal breast epithelial cells. NTRK3 encodes the neurotrophic tyrosine receptor kinase 3, a tyrosine kinase that is preferentially expressed in the central nervous system, where it is involved in the growth, development and survival of neuronal cells.33 NTRK3 protein is not expressed in normal breast epithelial cells. Given that the translocation t(12;15) leads to a fusion transcript, the transcriptional regulation of which is determined by the ETV6 gene promoter, it is ubiquitously expressed in SBC cells. This fusion protein has been shown to activate the Ras–Mek1 and PI3K–Akt pathways, which are associated with breast epithelial cell proliferation and survival.34
Although the ETV6–NTRK3 fusion gene has also been described in congenital fibrosarcoma,35 cellular mesoblastic nephroma36 and adult acute myeloid leukaemia,37 in the context of breast malignancies, this fusion gene seems to be restricted to SBC. Makretsov et al38 found no t(12;15)(p13;q25) translocation in 202 non-secretory breast carcinomas, whereas Letessier et al39 found five ETV6 gene rearrangements in non-secretory breast carcinomas, none of which showed an ETV6–NTRK3 fusion product. We have recently shown that acinic cell carcinomas, a subgroup of low-grade breast malignancies once believed to be related to SBCs, do not harbour the t(12;15)(p13;q25) translocation.40 In a way akin to soft tissue tumours and haematological malignancies harbouring recurrent balanced chromosomal translocations, SBCs display a rather simple karyotype. Diallo et al9 analysed 13 cases of SBC, 12 of which harboured the t(12;15) translocation, with comparative genomic hybridisation (CGH) and showed an average of 2.0 chromosomal aberrations/case. Recurrent gains of 8q (37.5%) and 1q (25%) and losses of 22q (25%) were found.9
Here we report on a unique case of SBC with a duplication of the t(12;15)(p13;q25) translocation involving ETV6–NTRK3, a feature that has not previously been observed in any of the SBCs reported in the literature.38 The duplication was identified following fluorescence in situ hybridisation analysis of the ETV6–NTRK3 translocation in a series of nine SBCs (data not shown). We also performed a detailed analysis of the genomic copy number aberrations associated with this phenomenon using tiling path microarray comparative genomic hybridisation (aCGH).
MATERIALS AND METHODS
An 82-year-old woman presented with a 2 cm lump in her right breast. A core biopsy was performed, which revealed a lesion composed of cells with ovoid nuclei harbouring discrete nucleoli. Their cytoplasm was either abundant, granular and pink or vacuolated, containing either multiple coalescent vacuoles or a large intracytoplasmic lumina. In these intracytoplasmic lumina, dense, deeply eosinophilic secretion was often found. Neoplastic cells were arranged in tubular and/or glandular, papillary and microcystic structures, containing densely eosinophilic PAS-positive diastase digestion-resistant secretion, conferring an appearance that vaguely resembled that of thyroid follicular structures (fig 1). A diagnosis of malignancy consistent with secretory carcinoma of the breast was made, and the patient underwent a wide local excision followed by radiotherapy. Gross inspection of the wide local excision revealed a lesion measuring 3.5×3.5×1.0 cm, which displayed a well-demarcated, solid and cystic, grey–white lesion, with cysts filled with yellow–brown mucoid material. Histological examination confirmed the diagnosis of secretory carcinoma of the breast, histological grade 1. No areas of vascular invasion were found. Immunohistochemical analysis was performed with antibodies against oestrogen receptor (ER, 1D5, 1:40; Dako, Glostrup, Denmark), progesterone receptor (PgR, PgR636, 1:150; Dako), androgen receptor (AR, AR441; 1:50, Dako, Glostrup, Denmark), HER2 (polyclonal, 1:1000; Dako), epidermal growth factor receptor (EGFR, 31G7, 1:50; Zymed, San Francisco, California, USA), cytokeratin (Ck) 5/6 (D5/16 B4, 1:600, Chemicon, Temecula, California, USA), Ck 8/18 (NCL-L-5D3, 1:100, Novocastra, Newcastle, UK), Ck 14 (LL002, 1:50, Biogenix Laboratory, San Ramon, California, USA), Ck 17 (E3(1), 1:100, Dako), Ck 19 (RCK108, 1:50, Dako), S100 protein (polyclonal, 1:6000, Dako) and gross cystic disease fluid protein 15 (GCDFP-15, 23A3, 1:40, Novocastra). Neoplastic cells lacked ER, PR, HER2, AR, GCDFP15 and Ck 14 expression, focally expressed Ck 5/6 and 17, and were positive for EGFR (10% of cells displaying moderate-to-strong membrane staining). Neoplastic cells displayed nuclear and cytoplasmic S100 protein expression and were diffusely positive for Ck 8/18 and 19 (fig 1). Since the course of radiotherapy, the patient has remained alive without disease for the last 44 months.
ETV6–NTRK3 probes were produced according to a previously described protocol (12). Briefly, the bacterial artificial chromosome (BAC) DNA was extracted using the QIAGEN Plasmid Mini Kit (QIAGEN, Crawley, West Sussex, UK) according to the manufacturer’s instructions, followed by amplification using the GenomiPhi Whole Genome Amplification Kit (WGA kit; GE Healthcare, Chalfont, Bucks, UK) according to the manufacturer’s instructions. All BAC clones used in this study were obtained from the BACPAC Resources Centre at the Children’s Hospital Oakland Research Institute. To validate the BACs and their genomic localisation, we performed BAC end sequencing and FISH mapping, respectively, as previously described.41
In this study, four sets of dual-colour probes were used to investigate the presence of the t(12;15) translocation in the tumour sample as described by Makretsov et al.38 The probes (table 1 and fig 2) were as follows: the ETV6 split-apart probe on chromosome 12 (probe AB), a NTRK3 split-apart probe on chromosome 15 (probe CD), the t(12;15) translocation fusion probes (probes AC and BD). The ETV6 split-apart probe contained a telomeric region to ETV6 on 12p (probe A, BAC clones RP11-434C1 and RP11-407P10) and was labelled with biotin, whereas the centromeric region to ETV6 (probe B, BAC clones RP11-525I3 and RP11-267J23) was labelled with digoxigenin. The NTRK3 split-apart probe consisted of two probes, the first being centromeric to NTRK3 (probe C, BAC RP11-114I9 and RP11-730G13 labelled with digoxigenin) and the second probe being telomeric to NTRK3 on 15q (probe D, BAC clones RP11-247E14 and RP11-893E1 labelled with biotin). In addition, we used probe E which comprised CEP12 (labelled with biotin) and CEP15 (labelled with digoxigenin) to asses the copy number of chromosomes 12 and 15 (kindly provided by Janet Shipley, Institute of Cancer Research, Sutton, UK).
Further validation of the ETV6 split-apart probe was performed using the ETV6 FISH DNA Probe Split Signal (Dako), which is composed of two ready-to-use FISH DNA probes: a Texas Red-labelled DNA probe (“ETV6-Upstream”), covering 264 kb telomeric to the ETV6 breakpoint cluster region, and a fluorescein-labelled DNA probe (“ETV6-Downstream”) covering 483 kb centromeric to the ETV6 breakpoint cluster region.
Representative histological sections were cut at 2 μm, mounted on polylysine-coated slides and subjected to FISH analysis as previously described.41 Briefly, dewaxed sections were heated with chromogenic in situ hybridisation (CISH) pretreatment buffer (SPOT-light tissue pretreatment kit; Zymed) on a hotplate for 15 min at ⩾98°C. The treated sections were then washed with water before being digested with pepsin (Zymed) for 5 min 30 s. Subsequently, sections were washed with water, dehydrated with graded ethanol and air-dried. Probes were resuspended in the CISH hybridisation buffer (60% deionised formamide, 12% dextran sulphate, 2.4×saline/sodium citrate SSC/0.14 mM EDTA, pH 8.0, 0.4 mg/ml salmon sperm DNA) and applied to the tissue section. The edges of the coverslip were sealed with rubber cement. Tissue sections and probes were denatured by placing the slides on a hot plate at 97°C for 7 min, before hybridisation overnight in a humidified chamber at 37°C. Coverslips were removed and slides were rinsed in 0.5×SSC at room temperature and then washed in 0.5×SSC for 7 min at 78°C. Non-specific protein binding was blocked with CAS-block (Zymed) for 10 min. The slides were then mounted with a mixture of 1 μl avidin-cyanine-3/μl anti-digoxigenin-fluorescein diluted 1:200 in phosphate-buffered saline/2% bovine serum albumin for 45 min in the dark at room temperature. The mixture was blotted off and the slides were washed with phosphate-buffered saline/0.025% Tween 20. The slides were counterstained with antifade medium containing 4′,6-diamidino-2-phenylindole. For the ETV6 Dako split-apart probes, FISH was performed as described by Reis-Filho et al.40 Images were collected sequentially in three channels (4′,6-diamidino-2-phenylindole, fluorescein isothiocyanate and cyanine-3) on a TCS SP2 confocal microscope (Leica, Milton Keynes, UK). For each experiment, at least 60 non-overlapping, interphase nuclei of neoplastic cells were analysed.
Four representative 8 μm-thick tumour tissue sections were cut, stained with nuclear fast red and treated with sodium thiocyanide as previously described.42 43 These sections were then microdissected using a sterile needle under a stereomicroscope (Olympus SZ61, Tokyo, Japan) to ensure a purity of neoplastic cells of >90%. Subsequently, DNA was extracted using DNeasy Kit according to the manufacturer’s protocol (QIAGEN). The purified DNA samples were then precipitated in ethanol and resuspended in Tris/EDTA buffer (pH 7.5). DNA yield and quality were assessed by picogreen assay and a multiplex quality-control PCR,43 respectively.
The aCGH platform used for this study was constructed at the Breakthrough Breast Cancer Research Centre and comprises ∼32 000 BAC clones tiled across the genome.44–46 The resolution of this aCGH platform is ∼50 kb.44–46 Labelling, hybridisation and washes were carried out as previously described.42 44–50 After hybridisation and washes, slides were scanned using an Axon 4000B scanner (Axon Instruments, Burlingame, California, USA), and images were processed using Genepix Pro 5.1 image analysis software (Axon Instruments). Log2 ratios were normalised for spatial and intensity-dependent biases using a two-dimensional loess regression. This left a final dataset of 31 423 clones with unambiguous mapping information according to the March 2006 build (hg18) of the human genome (http://www.ensembl.org). Data were smoothed using a local polynomial adaptive weights smoothing (aws) procedure for regression problems with additive errors.51 Thresholds for defining genomic gains and losses were obtained using data from unamplified female vs female and female vs male genomic DNA, as previously described.42 44 47 49 A categorical analysis was applied to clones on the array after classification as gain, loss or no change according to their smoothed log2 ratio values. Smoothed log2 ratio values <−0.16 were categorised as losses, those >0.16 as gains, and those in between as unchanged. Amplifications were defined as smoothed log2 ratio values >0.45. Data processing and analysis were carried out in R 2.0.1 (http://www.r-project.org/) and BioConductor 1.5 (http://www.bioconductor.org/), making extensive use of modified versions of the packages aCGH, marray and aws in particular.42 44–49 52–54
We used a combination of four probes to investigate the presence of the t(12;15)(p13;q25) ETV6–NTRK3 translocation as previously described38 (fig 2). FISH analysis revealed split-apart signals using the probe pairs AB and CD (fig 2A,B). ETV6 split-apart probes AB showed one normal copy of ETV6 (ie, juxtaposed signals) and two copies of the region telomeric to ETV6 (probe A). These results were independently confirmed using the ETV6 Dako split-apart probe (data not shown). NTRK3 split-apart probes (probe CD) showed one normal copy of NTRK3 and two copies of the region centromeric to NTRK3 (probe C) (table 2). The reciprocal translocation between chromosome 12 and chromosome 15, leading to the formation of der(15)t(12;15), was investigated by using the fusion probe 1 (a combination of probes A and C) and fusion probe 2 (a combination of probes B and D). FISH analysis using the fusion probe 1 revealed two juxtaposed red–green signals, suggesting two copies of the derivative chromosome der(15)t(12;15) and one additional copy of the region centromeric to NTRK3 (probe C) (fig 2C). FISH with the fusion probe 2 showed one copy of the region centromeric to ETV6 (probe B) and one copy of the region telomeric to NTRK3 (probe D) (fig 2D). Using these probes, no juxtaposed signals were observed, suggesting a lack of the derivative chromosome der(12)(12;15). FISH analysis using the centromeric probes of chromosome 12 (CEP12) and 15 (CEP15) revealed three copies of CEP15 in the modal population of neoplastic cells, but only one copy of CEP12 in the same population, suggesting a loss of CEP12 and a gain of CEP15. Taken together, our results suggest retention of one non-rearranged copy of ETV6 and NTRK3 genes, duplication of the derivative chromosome der(15)t(12;15) and loss of most of the derivative chromosome der(12)t(12;15) in tumour cells. Adjacent inflammatory cells, fibroblasts and endothelial cells displayed no rearranged signals and two copies of CEP12 and CEP15.
aCGH analysis revealed that this SBC harboured a simplex molecular karyotype,55 which is characterised by a few chromosomal aberrations involving whole chromosomes or chromosomal arms without the presence of amplifications. Gains of 4p16.3–p16.1, 5p15.33, 10q21, 10q26.3, 12p13.3–p13.31, 12p13.2, 15q11–q25.3, 15q26.2–15q26.3, 18q23, 19p13.3 and chromosome 16 (table 3 and fig 3) and losses of 6q24.1–q27, 12p13.2–q12, 15q2.3–q26.3, 14q22.3–q23.1 and 19q13.2 (table 3 and fig 3) were observed. Given the additional copy of the derivative chromosome der(15)t(12;15), aCGH revealed a gain of genetic material with the breakpoint mapping to exon 5 of the ETV6 gene on chromosome 12p13.2 (11.85 Mb) and to intron 5 of the NTRK3 gene on chromosome 15q25.1 (86.33 Mb). Our results also revealed a deletion of the region 86.30–97.21 Mb on chromosome 15 telomeric to the NTRK3 gene as well as a region 11.95–38.70 Mb on chromosome 12 centromeric to the ETV6 gene (fig 3). These findings corroborate the results of FISH analysis (ie, one copy of the probes B and D).
Here we describe a case of SBC with a duplication of the ETV6–NTRK3 translocation identified in a screen for the fusion gene in a series of special types of invasive breast cancers. To the best of our knowledge, this is the first example of an SBC with a duplication of a recurrent balanced chromosomal translocation and the first example of a solid epithelial malignancy harbouring a duplication of the ETV6–NTRK3 translocation. Interestingly, this case of SBC displayed a triple-negative phenotype (ie, lacked ER, PR and HER2) and expressed basal markers, including Cks 5/16 and 17 and EGFR. Interestingly, ETV6 has been shown to be expressed at significantly higher levels in basal-like breast cancers than non-basal-like tumours (analysis of data from Richardson et al56 using http://www.oncomine.org).
Taken together, our FISH and aCGH results are consistent with a gain of the derivative chromosome der(15)t(12;15) and a loss of genetic material on the derivative chromosome der(12)t(12;15), involving the fusion gene in this derivative chromosome (fig 4). On the basis of the results of the aCGH analysis, it is plausible that the duplication of the derivative chromosome der(15)t(12;15) may have resulted from a non-disjunctional mitosis. The partial loss of the derivative chromosome der(12)t(12;15) may have either preceded or, more likely, succeeded the duplication of the derivative chromosome der(15)t(12;15). Regardless of the actual mechanism giving rise to the duplication of ETV6–NTRK3, the modal population of tumour cells (>60%) displayed two copies of the derivative chromosome der(15)t(12;15) and lacked a substantial portion of the derivative chromosome der(12)t(12;15), suggesting that the duplication and loss either happened at the early stages of tumour development or conferred a growth advantage on tumour cells. Interestingly, in a way akin to the present case, a second copy of the derivative chromosome der(15)t(12;15) has been described in an acute myeloid leukaemia characterised by a similar ETV6–NTRK3 fusion gene.57 Furthermore, losses of der(12) have been described for congenital/infantile fibrosarcoma harbouring the t(12;15) translocation. Taken together, these findings suggest that loss of the derivative chromosome der(12)t(12;15) may result in deletion of elements on chromosome 12p that have tumour-suppressive activity or loss of 12p confers a direct growth advantage on neoplastic cells.
Owing to the duplication of the derivative chromosome der(15)t(12;15), aCGH analysis allowed the fine mapping of the breakpoints of the translocation: gain of the fusion gene partners ETV6–NTRK3 but with the loss of the region telomeric to NTRK3 (15q25.3–q26.3), and the region centromeric to ETV6 (12p13.2–q12), therefore confirming the FISH results. Although our aCGH analysis correctly mapped the breakpoint to the regions of ETV6 and NTRK3, these results should be interpreted with caution, as the resolution of the aCGH platform used in this study is limited to ∼50 kb.
Secretory breast cancers harbour the t(12;15)(p13;q25) translocation, which can be used as a diagnostic marker for these tumours.
Additional rearrangements involving the derivative chromosomes (ie, gain of der(15) and/or loss of der(12)) may confer additional growth advantage on neoplastic cells.
The presence of specific genetic changes in special types of breast cancer further corroborate the concept of a genotypic–phenotypic correlation in breast cancers.
Secretory breast cancers may have a triple-negative and basal-like phenotype.
Duplication of fusion genes is not a rare event in leukaemias and lymphomas.58–61 Duplication of the Philadelphia chromosome is associated with blastic transformation of chronic myelogenous leukaemia, resistance to imatinib, and overexpression of the BCR–ABL fusion gene.58–61 The duplication of ETV6–RUNX1 has been observed in five out of 56 cases of B-precursor acute lymphoblastic leukaemia.62 In contrast, duplication of fusion genes following chromosomal translocations is an uncommon phenomenon in solid malignancies. In a recent report by Attard et al,63 a fusion-gene duplication involving the TMPRSS2 and EGR genes, together with interstitial deletion of the 5′ sequence to EGR, was found in ∼6.6% (27/445) of prostate cancers and was associated with a poorer prognosis.
Our aCGH analysis also showed that this case of SBC has a “simplex” genomic profile55 (fig 3) with a small number of copy number changes across the genome (table 3 and fig 3). Hicks et al55 have recently shown that breast cancers with simplex genomic profiles are associated with low grade and improved patient outcome compared with tumours with complex/firestorm genomic profiles. Apart from the gains on 12p13.2–p13.31 (telomeric to ETV6) and 15q11–q25.3 (centromeric to NTRK3) and the losses of 12p13.2–q12 (centromeric to ETV6) and 15q25.3–q26.3 (telomeric to NTRK3), only a few additional changes not mapping to regions of copy number polymorphisms (http://projects.tcag.ca/variation/) were identified, namely gains of 10q21, 10q26.3 and chromosome 16 and losses of 6q24.1–q27. Our aCGH analysis showed normal copy numbers of 12q13–qtel, suggesting a possible translocation between der(6)t(6;12)(q24.1;q12), given that 12q12–qter does not have a centromere and theoretically would be unlikely to be retained if not rearranged with another chromosome (fig 4).
Previous studies have reported the loss of the long arm of chromosome 6 in breast cancer using loss of heterozygosity and CGH analysis.64–67 The oestrogen receptor gene 1 (ESR1) is located at 6q24.1–q27, which may account for the lack of ER expression in secretory breast cancer. However, it is well documented that loss of heterozygosity on the ESR1 locus does not affect the expression of ER in breast cancers.46 68 69 The region of 6q26 contains a chromosomal fragile site that encompasses the Parkin 2 gene (PARK2), and ∼6% of breast cancers have a breakpoint mapping to this gene.42 The gene myeloid/lymphoid or mixed-lineage leukaemia translocated 4 (AF-6/MLLT4), which maps to 6q27, encodes the afadin protein, which is widely expressed in normal epithelial cells. Loss of AF-6/MLLT4 protein expression has been found in 14.5% of breast cancers and is associated with poor prognosis in patients with deletions of this genomic locus.70
The pattern of genetic changes observed in the case presented here is in agreement with that reported by Diallo et al,9 who suggested that SBCs have a simple genomic profile and, unlike other types of low-grade breast cancer, do not harbour deletions of 16q. Interestingly, the chromosomes affected in the lesions described here differed from those previously described for secretory carcinomas (eg, no gains of 8q, 7q or 1q). However, given the low prevalence of chromosomal numerical aberrations in most secretory carcinomas reported to date and the presence of a duplication of the ETV6–NTRK3 translocation in the present case, one could speculate that, after the duplication of the balanced translocation, this tumour progressed through a distinct genetic pathway.
This work was funded by Breakthrough Breast Cancer.
Competing interests: None.
Ethics approval: Obtained.
Patient consent: Obtained.