Low-grade adenosquamous carcinoma (LGASC) is a rare type of metaplastic carcinoma of the breast (MBC) with an indolent clinical course. A few cases with high-grade transformation have been reported; however, their genetic alterations remain unclear. This study was designed to explore the somatic genetic characteristics of LGASC with transformation to high-grade MBC. Whole-genome sequencing analysis was performed on five MBCs from four patients, including one case with matching primary LGASC and lymph node metastatic tumor consisting of high-grade MBC with a predominant metaplastic squamous cell carcinoma pattern (MSC) that has progressed from LGASC, and three de novo MSC. Both LGASC and its associated MSC, unlike de novo MSC, had no TP53 mutation and tended to have fewer structural variants than de novo MSC. Both LGASC and its associated MSC had common mutations and copy number alterations, including GNAS R844C, which shows an increase in its allele frequency in MSC. These results indicated that LGASC and its associated MSC were ancestry clonal and that clonal selection occurred during progression. MSC associated with LGASC had additional pathogenic deletions of multiple tumor suppressor genes caused by, for example, SMAD4–DCC fusion. Reverse transcription polymerase chain reaction, followed by Sanger sequencing, confirmed this fusion transcript in both LGASC and its associated MSC; however, chimera proteins were not detected by Western blotting. SMAD4 protein expression had already decreased at the stage of LGASC, which may have contributed to the tumorigenesis of LGASC. In conclusion, not only LGASC but also its associated high-grade MBC may be genetically different from de novo high-grade MBC. The progression from LGASC to high-grade MBC may concern the concentration of the driver mutation caused by clonal selection and inactivation of tumor suppressor genes.
Article
Genetic analysis of low-grade adenosquamous carcinoma of the breast that progressed to high-grade metaplastic carcinoma
https://doi.org/10.21203/rs.3.rs-1982650/v1
This work is licensed under a CC BY 4.0 License
Version 1
posted
You are reading this latest preprint version
Metaplastic carcinoma of the breast (MBC) is a subtype of invasive breast carcinoma with differentiation to the squamous epithelium and mesenchymal component, accounting for 0.2–1% of all invasive breast carcinomas1. Most MBCs belong to triple-negative breast carcinoma. They usually have poor outcomes and lower response rates to conventional chemotherapy and a worse clinical outcome after chemotherapy than other forms of triple-negative breast carcinomas; however, obvious prognostic factors or predictive factors of therapeutic response remain unclear1. Meanwhile, MBCs are clinically, morphologically, and genetically heterogeneous tumors, and histologically low-grade tumors are rarely recognized. This low-grade group includes low-grade adenosquamous carcinoma (LGASC) and fibromatosis-like carcinoma and shows relatively favorable outcomes compared with high-grade tumors.
LGASC is a rare variant of MBC, characterized by well-developed glandular and tubular formation admixed with solid nests of squamous cells in a spindle-cell background1,2. Diagnosis using core biopsy is challenging because distinguishing between neoplastic and non-neoplastic glands and ducts is difficult. Although LGASC was suggested to be related to some benign breast proliferative lesions, such as intraductal papillary lesions, adenomyoepithelioma, and complex sclerosing lesions3–6, its pathogenesis is still unclear. Although LGASC generally has a favorable prognosis, few cases with a transition to high-grade MBC have been reported7–11. Molecular studies clarified that high-grade MBC frequently harbors TP53 mutation as observed in triple-negative invasive ductal carcinomas (no special type)12,13. One study revealed that MBC had more frequent mutations in PI3K-Akt pathway genes (57% vs. 22%), including PIK3CA (29%), PIK3R (11%), and PTEN (11%), and in Wnt pathway genes (51% vs. 28%), including FAT1 (37%), with pathway activation than triple-negative invasive ductal carcinoma (no special type)12. Other exome and targeted sequencing studies demonstrated frequent mutations in NF1 and KDM6A13,14. Meanwhile, molecular studies of LGASC have reported that LGASC had no TP53 mutation and that approximately half of LGASC cases had PIK3CA mutation and most cases had genetic alterations involved in the PI3K-Akt pathway, including PTEN4,7,15. However, the characteristics of the genetic alterations associated with the transition of LGASC to high-grade MBC have not been elucidated because of their rarity.
Therefore, this study was designed to explore the features of somatic genetic alterations in LGASC and high-grade MBC progressed from LGASC and compared them with those in high-grade de novo MBC.
Case selection
Frozen tissues of tumors and matching healthy skin tissues from surgically resected specimens were retrieved from the archives of the Kanagawa Cancer Center Biospecimen Center. We analyzed five MBC from four patients. The five MBCs included one primary LGASC (sample ID: LGASC) and high-grade MBC with a predominant metaplastic squamous cell carcinoma pattern (MSC) that progressed from LGASC and metastasized to the axillary lymph node (sample ID: LNMSC), and three de novo MSCs (sample ID: M2T, M3T, and M7T, respectively). The de novo MSCs were retrieved from frozen specimens for genetic analysis according to the sufficient cellularity of tumor cells. The remaining surgically resected tissues were fixed in 10% buffered formalin, embedded in paraffin, and stained with hematoxylin and eosin. The histology and histological grade of these tumors were evaluated according to the fifth edition of the WHO Classification of Tumours1. Clinical data, including germline mutations of breast cancer genes BRCA1 and BRCA2, stage, and outcomes, were retrieved from the medical records. The study was conducted according to the Declaration of Helsinki and was approved by the Ethics Committee of the Kanagawa Cancer Center (approval no. H28-240). Written informed consent for retrospective studies, including somatic and germline genetic analyses, as broad comprehensive consent, was obtained from the patients involved.
Immunohistochemistry
Immunohistochemical staining was performed on a section from the representative paraffin block of each tumor. Then, heat-induced antigen retrieval was performed. The following primary antibodies were used: estrogen receptor (ER) (SP1, Roche, Basel, Switzerland, pre-diluted), progesterone receptor （PgR）（1E2, Roche, pre-diluted）, human epidermal growth factor receptor 2 (HER2) (4B5, Roche, pre-diluted), p63 (4A4, Nichirei, Tokyo, Japan, pre-diluted), p40 (BC20, Nichirei, pre-diluted), SMA (1A4, Roche, pre-diluted), Ki-67 (MIB-1, Agilent, Santa Clara, CA, USA, pre-diluted), and SMAD4 (EP618Y, Abcam, Cambridge, UK, 1:100). All immunohistochemical procedures, except for p40 and SMAD4, were performed using the automatic staining machine VENTANA BenchMark ULTRA IHC/ISH system (Roche). ER, PgR, and HER2 were detected using the iVIEW DAB Detection Kit (Roche), and p63, SMA, and Ki-67 were detected using the ultraView Universal DAB Detection Kit (Roche). For p40, immunostaining was performed using HISTOSTAINER 48A (Nichirei), and its expression was detected using Histofine Simple Stain MAX-PO kit (Nichirei). For SMAD4, immunostaining was performed using BOND-Ⅲ (Leica Biosystems, Nussloch, Germany). Positive and negative controls were used. ER, PgR, and HER2 were evaluated according to the current American Society of Clinical Oncology/College of American Pathologists guidelines16,17.
Whole-genome sequencing
Genomic DNA was extracted from frozen tissues of each of the five MBCs and four matched healthy skin tissues using a standard method with protease K (PK) digestion. Briefly, the frozen tissues were minced and incubated at 65ºC overnight in 1-mg/mL PK solution, containing 10-mM Tris-HCl buffer (pH 7.8–8.0) (Thermo-Fisher Scientific, Inc., Waltham, MA), 0.4% (w/v) sodium dodecyl sulfate (NIPPON GENE CO., LTD, Tokyo, Japan), 150-mM NaCl, and 10-mM ethylenediaminetetraacetic acid (Thermo-Fisher Scientific, Inc.). The solution was extracted in 25:24:1 phenol/chloroform/isoamyl alcohol (phenol-CIAA) solution (FUJIFILM Wako Chemicals) several times, and finally, DNA was precipitated from the aqueous phase with ethanol and collected by centrifugation. DNA was once dissolved in autoclaved distilled water (dH2O) and treated with 10 µg/mL of a DNase-free RNase (Pure Link RNase A, Thermo-Fisher Scientific) at 37°C for 30 min. The reaction was treated once with phenol-CIAA, and DNA was obtained as described above. The double-stranded DNA was quantified with Qubit 2 (Thermo-Fisher Scientific), and DNA purity was evaluated using NanoPhotometer (Implen, Munich, Germany) as A260/A280 and A260/230 optical density ratios. WGS was entrusted to Genewiz (Shinagawa-ku, Tokyo, Japan) for primary data analysis. Briefly, sequencing libraries were constructed from each 100 ng of the extracted DNA using TruSeq Nano DNA Library Prep kit (Illumina, Hayward, CA) and sequenced using a standard 150-bp paired-end read protocol. The obtained sequence data were analyzed using the Genomon 2 DNA analysis pipeline (https://github.com/Genomon-Project) at the Human Genome Center, the Institute of Medical Science, the University of Tokyo (Tokyo, Japan).
Analysis of single-nucleotide variants and short insertions and deletions (indel)
Single-nucleotide variants (SNVs) with minimum depth ≥8, base quality ≥15, variant read ≥4, P <0.01 (Fisher’s exact test), and variant allele frequency ≥0.02 in the tumor were first selected for analysis. SNVs in protein coding regions that resulted in changes in amino acid sequences were further evaluated to identify pathogenic or likely pathogenic ones by interrogating against the Catalogue of Somatic Mutations in Cancer (COSMIC) (https://cancer.sanger.ac.uk/cosmic), the 1000 Genomes Project (https://international genome.org), dbSNPs (NCBI, NIH; https://www.ncbi.nlm.nih.gov/snp/), Functional Analysis through Hidden Markov Models (v2.3) (http://fathmm.biocompute.org.uk), ClinVar (NCBI, NIH; https://www.ncbi.nlm.nih.gov/clinvar/), and OncoKB (Memorial Sloan Kettering Cancer Center; https://www.oncokb.org/) databases. Literature searches were also conducted to recognize pathogenic or likely pathogenic SNVs. Breast cancer driver genes were finally determined in compliance with the list that appeared in the report by Michailidow et al.18.
Estimation of structural variations and DNA copy number variants
Large insertions, deletions, tandem duplications, and inversions, with tumor allele frequency ≥0.07, depth thresh fold ≥10, control depth ≥10, and inversion size threshold >1000 in the tumor were considered as structural variation (SV) candidates. Cellularity was measured using the R package Sequenza (https://sequenzatools.bitbucket.io). To examine the functional bias of the genes where the SVs breakpoints are localized, a clustering analysis of the gene subset on the SVs breakpoints was performed using DAVID 6.8 (https://david.ncifcrf.gov/home.jsp).
Copy number variant (CNV) profiling was performed using DNAcopy version 1.56.0. (https://bioconductor.org/packages/release/bioc/html/DNAcopy.html), an R/Bioconductor package. We identified significantly amplified or deleted regions of genome using GISTIC 2.019.
To detect chromothripsis from WGS data, we used ShatterSeek v1.1 (https:// github.com/parklab/ShatterSeek), an R package, and applied previously established criteria20
Evaluation of mutational patterns and signatures
MutationalPatterns version 2.0.0 (https://bioconductor.org/packages/release/bioc/html/MutationalPatterns.html), an R/Bioconductor-based package developed by Blokzijl et al.21, was used to characterize the patterns in base substitutions, and mutational signatures were identified through WGS.
Reverse transcription-polymerase chain reaction
The existence of SMAD4–DCC fusion transcript was evaluated in LGASC and LNMSC using reverse transcription-polymerase chain reaction (RT-PCR). Briefly, total RNA was extracted from frozen tissues of LGASC and LNMSC using TRIzol Reagent (Invitrogen, Waltham, MA) according to the manufacturer’s instructions. Then, total RNA was reverse transcribed into cDNA using SuperScript IV VILO Master Mix (Invitrogen) and subjected to PCR amplification using PrimeSTAR HS DNA polymerase (Takara, Kyoto, Japan). The forward PCR primer was set in the exon 3 of SMAD4 with the nucleotide sequence 5'-CTACGAACGAGTTGTATCACC-3' (SMAD4-forward), and the reverse primer was set in the exon 3 of DCC with the nucleotide sequence 5'-ACTTGAGTAGCACTGTGTCTC-3' (DCC-reverse). The amplified fragments that corresponded to the predicted molecular size were subcloned into pCR®︎-TOPO®︎ plasmid vector, and the nucleotide sequence was determined by Sanger sequencing using ABI PRISM 3130xl Genetic Analyzer (Thermo-Fisher Scientific, Inc.).
Western blotting
To confirm the existence of SMAD4–DCC chimeric protein, Western blotting was performed using the NuPAGE 4%–12% gradient Bis-Tris Protein Gel system (Thermo-Fisher Scientific, Inc.) with MOPS running buffer (Thermo-Fisher Scientific, Inc.). Proteins were extracted from frozen tissues of LGASC and LNMSC using Cell Lysis Buffer (Cell Signaling Technology, Inc., Dancers, MA). To detect SMAD4–DCC fusion protein, we used the following antibodies: anti-SMAD4 rabbit monoclonal antibody (BLR133J, Bethyl Laboratories, Montgomery, TX, USA, 1:1,000) and anti-DCC mouse monoclonal antibody (A-1, Santa Cruz Biotechnology, Dallas, TX, USA, 1:1,000). Anti-vinculin (V9131, Sigma-Aldrich, Merck KGaA, Darmstadt, Germany, 1:10,000) was used as a protein loading control. Secondary antibody reaction was performed using peroxidase-conjugated anti-mouse IgG (NA931, Cytiva, Tokyo, Japan, 1:100,000) or anti-rabbit IgG (NA934, Cytiva, 1:100,000). Detection was performed using the ImmunoStar LD-enhanced chemiluminescence detection reagent (FUJIFILM Wako Chemicals, Osaka, Japan).
Statistical analysis
Genetic data were compared using the Mann–Whitney U-test. Cosine similarity was used to determine the relative contribution of the 30 COSMIC mutational signatures (Mutational Signatures, version 2, https://cancer.sanger.ac.uk/signatures/signatures_v2/) in each sample. P values of less than 0.05 were used to denote statistical significance. All statistical analyses were performed using R, version 4.0.2.
Case presentation
A 45-year-old woman consulted the Kanagawa Cancer Canter to examine a palpable right breast mass of approximately 3.5 cm in diameter at the right upper central area with approximately 2-year history. Ultrasonography (US) identified an irregular isoechoic lesion, including a high echoic spot, surrounded by a focal peripheral low echoic area. US also found an axillary mass with a diameter of 1.0 cm with an inner echoic level resembling that of the breast mass. The histological diagnosis based on US-guided core-needle biopsy (CNB) and open biopsy of the breast mass was mastitis, and the patient began taking steroids after the denial of systemic infection. Consequently, the breast mass became smaller, and the axillary mass almost disappeared. However, 14 months after the initial CNB, both the breast and axillary masses started to grow rapidly. The axillary mass was resected, and the diagnosis was carcinoma, which was consistent with a breast cancer metastasis. As a result, the patient underwent Halsted mastectomy and axillary lymph node dissection. We made a final histopathological diagnosis of mixed metaplastic carcinoma (LGASC and MSC with a spindle cell carcinoma component). One month after surgical resection, skin, liver, and lymph node metastases appeared. Chemotherapy did not succeed, and the patient died of the disease 9 months after surgery.
Pathological findings
The initial CNB and open biopsy specimens showed irregular dilation of the ducts with secretion in the lumen and moderate chronic inflammation in the stroma. The nuclear atypia of the duct epithelium was bland. The immunohistochemical study identified p40-positive cells existing discontinuously in the periphery of the ducts (Figs. 1A and 1B). Gross examination of the mastectomy specimen showed a 6.0 × 5.5 × 4.5 cm ill-defined mass beneath the nipple, invading the dermis, nipple, and greater pectoral muscle. Cross-sections showed small cysts that contained mucinous material. Additionally, a 5.0 × 3.5 × 4.5 cm axillar mass was observed. Histologically, similar to what was observed on CNB, the superficial area of the breast tumor consisted of infiltrative glandular structures filled with abundant secretory materials in the lumens, resembling a microcystic pattern of secretory carcinoma (Fig. 1C). Mucin-containing cells were limited. Nuclear atypia was mild. A few tumor nests revealed squamous differentiation (Fig. 1D). Some bland spindle cells existed in the background of the glands and nests. The immunohistochemical study showed p40-positive cells discontinuously around the atypical glands and at the foci of squamous differentiation. Moreover, the p40-positive cells were partially SMA-positive, suggesting that these cells had mainly the basal cell phenotype and partially had the myoepithelial cell phenotype. The spindle cells surrounding the tumor were p40-negative and weakly SMA-positive. In the deep area, the tumor displayed a transition to high-grade MBC with clearer squamous cell differentiation, representing MSC with a high-grade adenosquamous carcinoma pattern and partially spindle cell morphology (Fig. 1E). Moderately inflammatory infiltrates, primarily lymphocytes and plasma cells, were found in the tumor. The axillary mass was a lymph node metastasis of the high-grade component of the breast tumor with extra-nodal invasion. These lesions consisted of MSC containing more spindle cell components than the deep area of the breast tumor (Fig. 1F). Partially, large tumor nests of pure squamous cell carcinoma with keratinization were observed (Fig. 1G). Some metastatic foci in the dissected lymph nodes included the metastatic LGASC component (Fig. 1H). Skin metastasis consisted of pure MSC. ER and PgR were negative, and the HER2 score was 1 + in tumor cells of both primary LGASC and MSC. The clinicopathological characteristics of the other three de novo MSCs from three patients are summarized in Table 1.
Somatic mutations and copy number variants
We performed WGS for the primary LGASC and the associated MSC component of the axillary lymph node metastasis (LNMSC), and three de novo MSC cases (sample ID: M2T, M3T, and M7T). The average sequencing coverage was 33.8 for the tumor samples and 34.3 for the healthy skin samples. The tumor cellularity of LGASC, LNMSC, M2T, M3T, and M7T estimated by Sequenza (https://sequenzatools.bitbucket.io) was 0.39, 0.18, 0.36, 0.38, and 0.39, respectively. We found that all reads aligned properly to the hg19 human genome reference sequence, thereby demonstrating the accuracy and reliability of the WGS performed. All somatic mutations satisfying the criterion set are listed in Supplementary Tables 1–5.
The median number of whole somatic mutations, such as SNVs and indels, and nonsynonymous mutations per sample was 2233 (range, 275–5167) and 30 (range, 13–93), respectively. Nonsynonymous mutations were equivalent to 1.8% of the entire mutations. The number of somatic mutations found in five samples was not statistically different from substitutions and insertions/deletions of the five MBCs from a previous report (median, 3058; range, 1676–13120; p = 0.2506 by the Mann–Whitney U-test)22. The nonsynonymous mutation count in these samples was not statistically different from that of the spindle or squamous MBC demonstrated in a previous report (median, 60; range, 11–223; p = 0.3195 by the Mann–Whitney U-test)12. However, the number of identified whole somatic mutations in LGASC tended to be lesser than that of high-grade MBCs.
The recurrent somatic mutations are shown in Fig. 2. Five nonsynonymous mutations in GNAS, SOAT2, SIPA1L3, LMTK3, and DHX34 were common between LGASC and LNMSC, and only the GNAS mutation (c.C2530T:p.Arg844Cys) was considered a pathogenic variant, which is known as a mutation hotspot of this gene23. The mutated GNAS allele frequencies in LGASC and LNMSC were 0.075 and 0.164, respectively. Two of the three de novo MSCs harbored pathogenic TP53 mutations. PIK3CA was also mutated in two of the three de novo MSCs, and one (M2T) had both mutations. Additionally, M7T showed other breast cancer driver gene mutations, including ERBB2 (c.G2329T:p.V777L) and HRAS (c.G34A:p.G12S), which are well-known hotspot mutations (Supplementary Table 5). A recurrent non-coding SNV was detected in an intron of NBPF25P (chromosome 1, 148557041 C > A), which was an ncRNA, in M3T and M7T (Supplementary Tables 4 and 5). No pathogenic mutations interpreted as homologous recombinational repair (HRR) genes annotated by gene ontology (http://amigo.geneontology.org/amigo/landing) were identified, except for TP53 mutations.
All SVs identified are shown in Supplementary Tables 6–10 and schematically summarized in Fig. 3A. The median total number of SVs identified in LGASC, LNMSC, and the three de novo MSCs was 107 (range, 9–250), which was not statistically different from the rearrangements of the five MBCs in a previous report (median, 113; range, 21–428; p = 0.4647 by the Mann–Whitney U-test)22. The genes at the breakpoint of the SVs were enriched with that encoding ubiquitin protein ligase by analysis using DAVID clustering analysis. The genes included in this cluster were RNF144A, TMEM189-UBE2V1, RABGEF1, UBE2V1, AMFR, SMURF1, TMEM189, and UBE2R2. We defined the genes located at the breakpoint of the SVs and those between the breakpoints of deletion as genes affected by SVs. Genes affected by SVs for which deletion or tandem duplication or fusion was curated to be oncogenic or likely oncogenic in OncoKB are listed in Fig. 3B. The genes included 13 known breast cancer driver genes, namely, STAG2, SMAD4, PTEN, PRDM1, PMS2, NF2, NF1, KMT2D, KMT2C, CIC, CDKN2B, CDKN1B, and BRCA1. BRCA1 is also an HRR gene. The number of SVs in LGASC and LNMSC was smaller than that of the three de novo MSCs (9 and 12 vs. 124, 250, and 107, respectively) in this study; however, the difference was not statistically significant (p = 0.2 by the Mann–Whitney U-test). The most frequently observed SV in LGASC was translocation, and that of the LNMSC was deletion. LNMSC contained deletions in multiple tumor suppressor genes, such as LATS1, KMT2D, and BTG1, compared with LGASC, which had no such SVs. These deletions were interpreted as probably heterozygous deduced from the allele frequency and tumor cellularity (18%, estimated by Sequenza). At chromosome 18q (chr18q), LGASC harbored 33bp deletion in intron 1 of SMAD4, whereas LNMSC harbored a large deletion spanning from chr18:48,579,729 in intron 4 of SMAD4 to chr18:50,300,003 in intron 2 of DCC (hg19), which may generate a SMAD4–DCC fusion gene (Fig. 3C). When confirming the genomic data using the Integrative Genome Viewer, only few reads corresponding to the same SMAD4–DCC fusion and deletion of TNFAIP and LATS1 were also found in LGASC. The most frequent SV in the three de novo MSCs was deletion. Whole gene deletion of RELN was observed in two of the three de novo MSCs, M2T and M7T. Although secretory and mucoepidermoid carcinomas were histologically considered differential diagnoses of MSC associated with LGASC, the reported ETV6-NTRK3 or CRTC1-MAML2 fusion genes in these tumors were not identified in our cases.
LGASC and LNMSC shared some CNVs; however, each harbored unique CNVs. In five MBCs, GISTIC peaks of copy number gain were observed at 8q24.3 (Supplementary Fig. 1), whereas no significant GISTIC peaks of copy number loss were identified.
Since de novo MSCs, especially M2T and M7T, had a large number of SVs concentrated in a few chromosomes, we interrogated the occurrence of chromothripsis in three de novo MSCs using an R package ShatterSeek. However, no genetic events satisfying the criteria of chromothripsis were found in any of the specimens.
Analysis of SMAD4–DCC fusion and SMAD4 expression in low-grade adenosquamous carcinoma and its associated metaplastic squamous cell carcinoma
WGS analysis revealed a possibility that SMAD4–DCC fusion transcript and protein were generated in LNMSC (Fig. 4A). Therefore, we first performed RT-PCR to detect a predicted fusion transcript, joining exon 4 of SMAD4 and exon 3 of DCC. The forward PCR primer was set in exon 3 of SMAD4 because exon 4 was very short with only 30 base pairs. Two clear fragments were amplified by the designed RT-PCR (Fig. 4B), and the determination of the nucleotide sequence revealed that both fragments were derived from SMAD4–DCC fusion transcripts. The larger product was 128 base pairs long containing exon 3 (downstream of the PCR primer) and exon 4 of SMAD4 and exon 3 of DCC (upstream of the PCR primer), and the smaller product was 98 base pairs long without exon4 of SMAD4. Both transcripts were in frame for both wild-type SMAD4 and DCC polypeptides (Fig. 4C). However, the predicted SMAD4–DCC fusion protein was not identified by Western blotting analysis using an anti-SMAD4 antibody that recognizes the N-terminus of SMAD4 and an anti-DCC antibody that recognizes the C-terminal part of DCC, whereas bands of SMAD4 and DCC corresponding to the wild-type protein were recognized in both LGASC and LNMSC (Fig. 4D). The immunohistochemical study using an anti-SMAD4 antibody that recognizes the C-terminus of SMAD4 was conducted to examine SMAD4 distribution in tissues. In LGASC, SMAD4 was weakly positive in a few nuclei of glandular epithelial cells compared with that in the nearby non-neoplastic breast epithelial cells, whereas it was mostly negative in the MSC associated with LGASC (Fig. 4E).
Mutational patterns and mutational signatures
The mutational patterns for each of the five MBCs were demonstrated according to the relative contribution of the identified mutation numbers for all 96 trinucleotide changes (Fig. 5A). As the SNV patterns, C > T/G > A transition was the most frequent pattern, followed by the C > G/G > C transversion, which was consistent with a previously reported pattern of breast cancer24. The relative contribution of the 30 well-known COSMIC mutational signatures to the mutational profile of each MBC sample revealed that the most dominant mutational signature was signature 3 (associated with BRCA1/2 mutations), which was consistent with a previous report (Fig. 5B)12; however, the relative contribution of the signature was variable in each sample. These mutational signatures were further evaluated using the relative contribution method using an R package MutationalPatterns (Supplementary Fig. 2). The five MBCs were hierarchically clustered, with cases in which histological squamous cell carcinoma components accounted for 90% or more (corresponding to M2T and M7T) and cases in which less than 90% (corresponding to LGASC, LNMSC, and M3T) clustered separately. LNMSC did not segregate with de novo MSC in hierarchical clustering.
In this study, we demonstrated the genetic profile of an LGASC that converted to high-grade MBC with clonal concentration and accumulation of loss of function mutations in multiple tumor suppressor genes. The converted high-grade MBC was genetically different with de novo high-grade MBC, suggesting a new possibility for the development of breast cancer.
The progression from LGASC to high-grade MBC is a rare phenomenon, with at least 10 cases reported. Among the cases with known histology of high-grade components, spindle cell carcinoma is the most common, and squamous cell carcinoma is included in two cases7–11. Of the 10 cases, prognostic information was available in three cases, and lymph node metastasis, distant metastasis, or cancer deaths were not reported, with their follow periods ranging from less than 6 months to 11 years7,11. There was one molecular study about transformation. Geyer et al. examined the CNV of one LGASC case transitioning to high-grade MBC, consisting of malignant spindle cells, using microarray comparative genomic hybridization, followed by chromogenic in situ hybridization9. This LGASC case displayed a rather simple molecular karyotype, comprising gains of 6p, 7pq, and 8q and losses of 1p, 6p, 6q, 8p, and 9p, without amplifications. In contrast, an LGASC case with progression to high-grade MBC displayed a greater degree of genomic complexity, comprising gains of 1q, 5q, 7p, 8q, 12p, 14q, 16p, 16q, and 18pq and losses of 1p, 3q, 8p, 9p, 12q, 17p, 17q, 22q, and Xpq, with high-level amplification of 7p11.2, encompassing the EGFR gene, and 7q11.21. Chromogenic in situ hybridization for the EGFR gene showed that not only epithelial cells but also few spindle cells in the stromal component harbored EGFR amplification, and spindle cells harboring EGFR amplification were found in both the low- and high-grade areas of this tumor.
In our case, multiple SNVs, SVs, and CNVs common to both LGASC and LNMSC proved that LGASC and its associated MSC have a common clonal ancestry. Additionally, LGASC and LNMSC had unique gene alterations, respectively. It means that LGASC with high-grade transformation is constituted by a mixture of multiple tumor clones, which had acquired diverse mutations as tumor progression. Many more SNVs were detected in LNMSC than in LGASC; however, no additional pathogenic SNVs were found in LNMSC. Furthermore, LNMSC had less CNVs than LGASC. In this respect, SNV and CNV are most likely not the leading cause of progression from LGASC to high-grade MBC in our case. LNMSC displayed additional deletions of tumor suppressor genes, such as LATS1, KMT2D, and BTG1. LATS1 is a serine/threonine-protein kinase that is a regulator of the Hippo signal transduction pathway25. LATS1 is believed to be a tumor suppressor that inactivates oncogene YAP1/TAZ, which mediates cellular survival pathways26. In breast cancer, Takahashi et al. demonstrated that 17/30 (56.7%) breast cancer had LATS1 promoter hypermethylation with lower mRNA levels. Decreased expression of LATS1 mRNA was significantly associated with large tumor sizes, high risk of lymph node metastasis, estrogen receptor and progesterone receptor negativity, and shorter relapse-free survival27. KMT2D (also known as MLL2) functions as an epigenetic modulator that methylates lysine residue 4 on the tail of histone H3 (H3K4)28. The deletion of KMT2D in human cells leads to genomic instability due to transcriptional stress29. Recently, Morcillo-Garcia et al.30 identified KMT2D mutation, and low transcriptomic levels of KMT2D were related to poor relapse-free survival of breast cancer using the METABRIC database. BTG1 is a member of an anti-proliferative gene family that regulates cell growth and differentiation31. Significantly lower levels of BTG1 mRNA were observed in breast cancer cell lines than in healthy cells with promoter hypermethylation, suggesting that BTG1 expression through epigenetic suppression is involved in mammary carcinogenesis32. Recent studies have demonstrated that decreased BTG1 expression is correlated with aggressive biological behavior and worse overall survival in breast cancer 33. BTG1 inhibits cell growth by inducing cell cycle arrest and apoptosis in breast carcinoma cells, which revealed that BTG1 plays a tumor-suppressive role in breast cancer34. These results support the notion that clonal selection, with tumor cell enrichment that has GNAS gene alteration, with stepwise inactivation of tumor suppressor genes rather than addition of driver gene activation had driven the progression to high-grade MBC from LGASC (Fig. 6). The high-grade MBC derived from LGASC and de novo high-grade MBC have similar squamous morphology; however, unlike the latter, the former has less SVs and does not have TP53 loss of function mutation. There are several speculations regarding the cellular origin of metaplastic carcinoma, such as dedifferentiation from conventional adenocarcinoma or development from basal-like stem cells. Metaplastic carcinoma with fewer SVs and lack of TP53 mutation could be derived from the low-grade variant of MBC as the presented case.
In all five MBCs, signature 3 was a predominant mutational signature, which is associated with the failure of DNA double-strand break-repair by homologous recombination and strongly associated with germline and somatic BRCA1 and BRCA2 mutations in breast, pancreatic, and ovarian cancers22,35. However, in all five MBCs, only M3T had BRCA1 pathogenic somatic mutations, but no germline BRCA1/2 pathogenic mutations, were identified. According to the study by Weigelt et al., 30% of MBCs showed homozygous recombinant deficiency evaluated by large-scale state transition36. However, BRCA1/2 total exome sequencing analysis showed only two (6%) of likely pathogenic mutations in the BRCA2 gene in 35 MBCs12. These results indicate the possibility of other mechanisms of BRCA1/2 inactivation, such as BRCA1 promoter methylation, or abnormalities in other genes may also generate it22,35. Further studies with more samples are warranted to elucidate the mechanism of genomic instability, which could generate multiple SVs.
WGS analysis showed that two of the three de novo high-grade MBCs (M2T and M3T) had pathogenic TP53 mutations, whereas LGASC did not. This was consistent with the results of previous studies4,7,12,13,15. Previous studies demonstrated that approximately half of LGASC cases had PIK3CA mutation4,7,15; however, the LGASC case in this study did not have this alteration, whereas a GNAS hotspot SNV was indicated. GNAS encodes G protein-stimulating α subunit (Gsα), which functions as a mediator in the G protein-coupled receptor signaling pathway, which is associated with the activation of adenylyl cyclase and various cellular responses37. p.Arg201His or p.Arg201Cys (p.Arg844His or p.Arg844Cys when based on the NM_080425 sequence) disrupts intrinsic GTPase activity of Gsα, resulting in the constitutive activation of Gsα38. GNAS mutations are most commonly associated with appendiceal mucinous neoplasms, appendiceal mucinous adenocarcinomas, pancreatic intraductal papillary mucinous neoplasms, and invasive adenocarcinomas arising from intraductal papillary mucinous neoplasms. GNAS is one of the breast cancer driver genes with a somatic mutation frequency of 1.0% according to The Cancer Genome Atlas Pan-Cancer Atlas (https://www.cbioportal.org), and high-grade MBC has more GNAS mutations (2.9–6.7%)12,13. Both primary LGASC and LNMSC had the same pathogenic somatic GNAS mutation (p.Arg844Cys), which was observed in one case in a previous study with target sequencing of LGASC series7. Recently, two reports demonstrated that three of five ductal adenomas and one of 10 adenomyoepitheliomas of the breast harbored this mutation39,40. Multiple case series suggest that benign proliferative lesions, such as ductal adenoma and adenomyoepithelioma, are precursor lesions of LGASC3,5,6,41. Still, no such benign lesions were found in the background in our case. Regarding whether a benign proliferative lesion is a precursor to LGASC, a study involving numerous cases is needed. Komatsu et al. showed some pancreatic cancer cell lines with exogenous GNAS mutations that alter mucin gene expression, such as MUC2 and MUC5AC, which do not promote cell proliferation42. It is speculated that GNAS mutation might be related to the histology of abundant secretion in this case. There was no description whether the recently reported LGASC case with GNAS hotspot mutation has hypersecretory morphology7. The relationship between genetic alteration and the unusual rich secretory morphology of this case has remained unclear and thus needs more investigation with molecular biological research.
One of the most frequent loss of heterozygosity regions of breast cancer is 18q, and SMAD4 and DCC located at 18q21 are inactivating tumor suppressor gene candidates43. SMAD4 is a critical effector of the TGF-β signaling pathway. This SMAD-dependent pathway plays an important role in the activation of cell cycle arrest and apoptosis of epithelial cells and the maintenance of genomic integrity. Additionally, the TGF-β pathway also promotes epithelial–mesenchymal transition in SMAD4-dependent and SMAD4-independent manners44. SMAD4 frequently mutates in pancreatic and colorectal cancers, and the absence of or decreased SMAD4 expression has been found in various cancers. Although the function of SMAD4/TGF-β in tumor development is seemingly controversial in the literature45,46, Liu et al. showed that SMAD4 expression is significantly lower in breast cancer tissues than in healthy breast epithelia and that reduced SMAD4 protein expression in primary invasive breast cancer correlates with high grade and poor prognosis46. The inactivation of the TGF-β pathway usually occurs due to the biallelic inactivation of SMAD4; however, the homozygous deletion of SMAD4 is rare in breast cancer. DCC (DCC Netrin 1 receptor) is a transmembrane protein that mediates axon guidance and induces apoptosis in the absence of a ligand. This protein functions as a tumor suppressor and frequently shows loss of expression in colorectal cance47. Few homozygous co-deletions of DCC and SMAD4 were reported in breast cancer. DCC expression decreased in 45 of the 75 (60%) of breast cancer tissues, and this decrease was correlated with recurrence or death in the univariate analysis48. These results lead to the consideration that other mechanisms, such as epigenetic reaction or protein degradation, may decrease SMAD4 and DCC protein expression in breast cancer. We found a novel SMAD4–DCC fusion gene generated by deletion in LGASC and its associated MSC using WGS. Although SMAD4–DCC fusion transcripts were detected in both LGASC and LNMSC using RT-PCR, no fusion protein was identified using Western blotting. Possible causes of these results were that fusion transcripts might be not translated due to the instability of fusion transcripts and that fusion proteins might be easy to degrade even if produced. Although SMAD4 and DCC deletion was not considered a homozygous deletion from their allele frequency, an immunohistochemical study showed that SMAD4 was weakly positive in a small part of the glandular epithelium only in LGASC and almost negative in the cells of its associated MSC. These results inferred that SMAD4 protein expression has already been suppressed from the stage of LGASC for uncertain mechanisms, suggesting that this SMAD4 reduction contributes to the tumorigenesis of LGASC. The effect of 33bp deletion of SMAD4 intron 1 in LGASC was unlikely to be the mechanism because the location of this short deletion was far from the splice site. Possible causes of this reduced expression are epigenetic modification and protein degradation. Additionally, in MSC, haploinsufficiency and dominant negative effect can also be the cause49,50.
In conclusion, we demonstrated the genetic profile of an LGASC that progressed to MSC with GNAS hotspot mutation, without TP53 loss of function mutation or pathogenic PI3K signaling pathway-related gene alterations, indicating that the genetic characteristics of LGASC and its associated high-grade MBC are different from those of de novo high-grade MBC. Clonal selection with the concentration of driver gene mutations and tumor suppressor gene inactivation could explain a part of the mechanism of progression.
Acknowledgments
The authors thank the experts from whom external opinion was sought in this uncommon histological subtype of breast cancer. The authors also acknowledge the support received from the technical staff of the Kanagawa Cancer Center for their expert technical support and the members of the Clinical Sequencing Team at the Human Genome Center, Institute of Medical Science, The University of Tokyo, for their expert technical assistance and excellent discussion. The super-computing resource was provided by the Human Genome Center, Institute of Medical Science, The University of Tokyo (http://sc.hgc.jp/shirokane.html). Genomon 2 was provided by Kenichi Chiba, Ai Okada, and Yuichi Shiraishi. The authors would like to thank Enago (www.enago.jp) for the English language review.
Ethical approval and consent to participate
All procedures in this study involving human participants were performed according to the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. This study was approved by the Ethics Committee of the Kanagawa Cancer Center (approval no. H28-240). Individual consent for this retrospective analysis was withheld as comprehensive consent was obtained.
Author Contributions
KK, TX, and YM developed the study concept and design; KK, TX, RK, and YM contributed to the writing, review, and revision of the paper; KK, TX, RK, ES, KK, RY, KY, KY, YF, SM, SI, and YM contributed to the acquisition, analysis, and interpretation of data and statistical analysis; RK, TY, EY, KW, YO, SS, and TY provided technical and material support. All authors have read and approved the final version of this manuscript for submission.
Funding
This work was supported by JSPS KAKENHI Grant Numbers JP16H06277 and JP22H04923 (CoBiA) to YM.
Data Availability Statement
The datasets used and/or analyzed during this study are available from the corresponding author on reasonable requests.
Competing Interests
The authors declare no competing financial interests.
Conflict of interest
The authors declare no competing financial interests.
- Reis-Filho JS, Sotiriou C, Vincent-Salomon A, Rakha EA, Shin S, Gobbi H et al. Metaplastic Carcinoma. In: Allison KH et al., editors. World Health Organization Breast Tumours. Lyon: IARC; 2019. p. 134–138.
- McMullen ER, Zoumberos NA, Kleer CG. Metaplastic breast carcinoma: update on histopathology and molecular alterations. Arch Pathol Lab Med 143, 1492–1496 (2019).
- Denley H, Pinder SE, Tan PH, Sim CS, Brown R, Barker T et al. Metaplastic carcinoma of the breast arising within complex sclerosing lesion: a report of five cases. Histopathology 36, 203–209 (2000).
- Eberle CA, Piscuoglio S, Rakha EA., Ng CKY, Geyer FC, Edelweiss M et al. Infiltrating epitheliosis of the breast: characterization of histological features, immunophenotype and genomic profile. Histopathology 68, 1030–1039 (2016).
- Gobbi H, Simpson JF, Jensen RA, Olson SJ, Page DL. Metaplastic spindle cell breast tumors arising within papillomas, complex sclerosing lesions, and nipple adenomas. Mod Pathol 16, 893–901 (2003).
- Van Hoeven KH, Drudis T, Cranor ML, Erlandson RA, Rosen PP. Low-grade adenosquamous carcinoma of the breast. A clinocopathologic study of 32 cases with ultrastructural analysis. Am J Surg Pathol 17, 248–258 (1993).
- Bataillon G, Fuhrmann L, Girard E, Menet E, Laé M, Capovilla M et al. High rate of PIK3CA mutations but no TP53 mutations in low-grade adenosquamous carcinoma of the breast. Histopathology 73, 273–283 (2018).
- Boecker W, Stenman G, Loening T, Andersson MK, Sinn H. P, Barth P et al. Differentiation and histogenesis of syringomatous tumour of the nipple and low-grade adenosquamous carcinoma: evidence for a common origin. Histopathology 65, 9–23 (2014).
- Geyer FC, Lambros MB, Natrajan R, Mehta R, Mackay A, Savage K et al. Genomic and immunohistochemical analysis of adenosquamous carcinoma of the breast. Mod Pathol 23, 951–960 (2010).
- Kawaguchi K, Shin SJ. Immunohistochemical staining characteristics of low-grade adenosquamous carcinoma of the breast. Am J Surg Pathol 36, 1009–1020 (2012).
- Rosen PP, Ernsberger D. Low-grade adenosquamous carcinoma. A variant of metaplastic mammary carcinoma. Am J Surg Pathol 11, 351–358 (1987).
- Ng CKY, Piscuoglio S, Geyer FC, Burke KA., Pareja F, Eberle CA et al. The Landscape of Somatic Genetic Alterations in Metaplastic Breast Carcinomas. Clin Cancer Res 23, 3859–3870 (2017).
- McCart Reed AE, Kalaw E, Nones K, Bettington M, Lim M, Bennett J et al. Phenotypic and molecular dissection of metaplastic breast cancer and the prognostic implications. J Pathol 247, 214–227 (2019).
- Dieci MV, Smutnà V, Scott V, Yin G, Xu R, Vielhet P et al. Whole exome sequencing of rare aggressive breast cancer histologies. Breast Cancer Res Treat. 156, 21–32 (2016).
- Wilsher MJ, Owens TW, Allcock RJ. Next generation sequencing of the nidus of early (adenosquamous proliferation rich) radial sclerosing lesions of the breast reveals evidence for a neoplastic precursor lesion. J Pathol Clin Res 3, 115–122 (2017).
- Allison KH., Hammond MEH, Dowsett M, McKernin SE, Carey LA, Fitzgibbons PL et al. Estrogen and progesterone receptor testing in breast cancer: ASCO/CAP guideline update. J Clin Oncol 38, 1346–1366 (2020).
- Wolff AC, Hammond MEH, Allison KH, Harvey BE, Mangu PB, Bartlett JMS et al. Human epidermal growth factor receptor 2 testing in breast cancer: American Society of Clinical Oncology/College of American Pathologists Clinical Practice guideline focused update. J Clin Oncol 36, 2105–2122 (2018).
- Michailidou K, Lindström S, Dennis J, Beesley J, Hui S, Kar S et al. Association analysis identifies 65 new breast cancer risk loci. Nature 551, 92–94 (2017).
- Mermel CH, Schumacher SE, Hill B, Meyerson ML, Beroukhim R, Getz G. GISTIC2.0 facilitates sensitive and confident localization of the targets of focal somatic copy-number alteration in human cancers. Genome Biol 12, R41 (2011).
- Cortés-Ciriano I, Lee JJ, Xi R, Jain D, Jung YL, Yang L, et al. Comprehensive analysis of chromothripsis in 2,658 human cancers using whole-genome sequencing. Nat Genet 52, 331–341 (2020).
- Blokzijl F, Janssen R, van Boxtel R, Cuppen E. MutationalPatterns: comprehensive genome-wide analysis of mutational processes. Genome Med 10, 33 (2018).
- Nik-Zainal S, Davies H, Staaf J, Ramakrishna M, Glodzik D, Zou X et al. Landscape of somatic mutations in 560 breast cancer whole-genome sequences. Nature 534, 47–54 (2016).
- Chang MT, Bhattarai TS, Schram AM, Bielski CM, Donoghue MTA, Jonsson P et al. Accelerating discovery of functional mutant alleles in cancer. Cancer Discov 8, 174–183 (2018).
- Greenman C, Stephens P, Smith R, Dalgliesh GL, Hunter C, Bignell G et al. Patterns of somatic mutation in human cancer genomes. Nature 446, 153–158 (2007).
- Johnson R, Halder G. The two faces of Hippo: targeting the Hippo pathway for regenerative medicine and cancer treatment. Nat Rev Drug Discov 13, 63–79 (2014).
- Hao Y, Chun A, Cheung K, Rashidi B, Yang X. Tumor suppressor LATS1 is a negative regulator of oncogene YAP. J Biol Chem 283, 5496–5509 (2008).
- Takahashi Y, Miyoshi Y, Takahata C, Irahara N, Taguchi, T, Tamaki Y et al. Down-regulation of LATS1 and LATS2 mRNA expression by promoter hypermethylation and its association with biologically aggressive phenotype in human breast cancers. Clin Cancer Res 11, 1380–1385 (2005).
- Rao RC, Dou Y. Hijacked in cancer: the KMT2 (MLL) family of methyltransferases. Nat Rev Cancer 15, 334–346 (2015).
- Kantidakis T, Saponaro M, Mitter R, Horswell S, Kranz A, Boeing S et al. Mutation of cancer driver MLL2 results in transcription stress and genome instability. Genes Dev 30, 408–420 (2016).
- Morcillo-Garcia S, Noblejas-Lopez MM, Nieto-Jimenez C, Perez-Peña J, Nuncia-Canatarero M, Győrffy B et al. Genetic mutational status of genes regulating epigenetics: Role of the histone methyltransferase KMT2D in triple negative breast tumors. PLoS One 14, e0209134 (2019).
- Winkler GS. The mammalian anti-proliferative BTG/Tob protein family. J Cell Physiol 222, 66–72 (2010).
- Woo HY, Do SI, Kim SH, Song SY, Kim HS. Promoter Methylation Down-regulates B-cell Translocation Gene 1 Expression in Breast Carcinoma. Anticancer Res 39, 5361–5367 (2019).
- Sheng SH, Zhao CM, Sun GG. BTG1 expression correlates with the pathogenesis and progression of breast carcinomas. Tumour Biol 35, 3317–3326 (2014).
- Zhu R, Zou ST, Wan JM, Li W, Li XL, Zhu W et al. BTG1 inhibits breast cancer cell growth through induction of cell cycle arrest and apoptosis. Oncol Rep 30, 2137–2144 (2013).
- Alexandrov LB, Nik-Zainal S, Wedge DC, Aparicio SA, Behjati S, Biankin AV et al. Signatures of mutational processes in human cancer. Nature 500, 415–421 (2013).
- Weigelt BN, CKY, Shen R, Popova T, Schizas M, Natrajan R et al. Metaplastic breast carcinomas display genomic and transcriptomic heterogeneity [corrected]. Mod Pathol 28, 340–351 (2015).
- Dhanasekaran DN. Transducing the signals: a G protein takes a new identity. Sci STKE 2006, pe31 (2006).
- Landis CA, Masters SB, Spada A, Pace AM, Bourne HR, Vallar L. GTPase inhibiting mutations activate the alpha chain of Gs and stimulate adenylyl cyclase in human pituitary tumours. Nature 340, 692–696 (1989).
- Lubin D, Toorens E, Zhang PJ, Jaffer S, Baraban E, Bleiweiss IJ et al. Adenomyoepitheliomas of the breast frequently harbor recurrent hotspot mutations in PIK3-AKT pathway-related genes and a subset show genetic similarity to salivary gland epithelial-myoepithelial carcinoma. Am J Surg Pathol 43, 1005–1013 (2019).
- Volckmar AL, Leichsenring J, Flechtenmacher C, Pfarr N, Siebolts U, Kirchner M et al. Tubular, lactating, and ductal adenomas are devoid of MED12 Exon2 mutations, and ductal adenomas show recurrent mutations in GNAS and the PI3K-AKT pathway. Genes Chromosomes Cancer 56, 11–17 (2017).
- Foschini MP, Pizzicannella G, Peterse JL, Eusebi V. Adenomyoepithelioma of the breast associated with low-grade adenosquamous and sarcomatoid carcinomas. Virchows Arch 427, 243–250 (1995).
- Komatsu H, Tanji E, Sakata N, Aoki T, Motoi F, Naitoh T et al. A GNAS mutation found in pancreatic intraductal papillary mucinous neoplasms induces drastic alterations of gene expression profiles with upregulation of mucin genes. PLoS One 9, e87875 (2014).
- Miller BJ, Wang D, Krahe R, Wright FA. Pooled analysis of loss of heterozygosity in breast cancer: a genome scan provides comparative evidence for multiple tumor suppressors and identifies novel candidate regions. Am J Hum Genet 73, 748–767 (2003).
- Ahmed S, Bradshaw AD, Gera S, Dewan MZ, Xu R. The TGF-β/Smad4 Signaling Pathway in Pancreatic Carcinogenesis and Its Clinical Significance. J Clin Med 6 (2017).
- Ikushima H, Miyazono K. TGFbeta signalling: a complex web in cancer progression. Nat Rev Cancer 10, 415–424 (2010).
- Liu N, Yu C, Shi Y, Jiang J, Liu Y. SMAD4 expression in breast ductal carcinoma correlates with prognosis. Oncol Lett 10, 1709–1715 (2015).
- Mehlen P, Fearon ER. Role of the dependence receptor DCC in colorectal cancer pathogenesis. J Clin Oncol 22, 3420–3428 (2004).
- Koren R, Dekel Y, Sherman E, Weissman Y, Dreznik Z, Klein B et al. The expression of DCC protein in female breast cancer. Breast Cancer Res Treat 80, 215–220 (2003).
- Zhang Y, Feng X, We R, Derynck R. Receptor-associated Mad homologues synergize as effectors of the TGF-beta response. Nature 383, 168–172 (1996).
- de Caestecker MP, Yahata T, Wang D, Parks WT, Huang S, Hill CS et al. The Smad4 activation domain (SAD) is a proline-rich, p300-dependent transcriptional activation domain. J Biol Chem 275, 2115–2122 (2000).
Table 1. Clinicopathological characteristics of a low-grade adenosquamous carcinoma with progression to high-grade metaplastic carcinoma and three de novo high-grade metaplastic carcinomas
CT, chemotherapy; RT, radiotherapy; NA, not applicable; DFS, disease free survival; OS, overall survival; DOD, dead of disease; NED, no evidence of disease; IBC-NST, invasive breast carcinoma of no special type; SqCC, squamous cell carcinoma; SpCC, spindle cell carcinoma; HG, histological grade according to reference 1. *Histological pattern is considered according to the WHO classification of tumours 5th edition for primary lesions and lymph node metastases. †It contains about 20% adenosquamous carcinoma pattern.
There is NO conflict of interest to disclose.