Molecular Cytogenomic Characterization of the Murine Breast Cancer Cell Lines C-127I, EMT6/P and TA3 Hauschka

Background: To test and introduce effective and less toxic breast cancer (BC) treatment strategies, animal models, including murine BC cell lines, are considered as perfect platforms. Strikingly, the knowledge on the genetic background of applied BC cell lines is often sparse though urgently necessary for their targeted and really justified application. Methods: In this study, we performed the first molecular cytogenetic characterization for three murine BC cell lines C-127I, EMT6/P and TA3 Hauschka. Besides fluorescence in situ hybridization-banding, array comparative genomic hybridization was also applied. Thus, overall, an in silico translation for the detected imbalances and chromosomal break events in the murine cell lines to the corresponding homologous imbalances in humans could be provided. The latter enabled a comparison of the murine cell line with human BC cytogenomics. Results: All three BC cell lines showed a rearranged karyotype at different stages of complexity, which can be interpreted carefully as reflectance of more or less advanced tumor stages. Conclusions: Accordingly, the C-127I cell line would represent the late stage BC while the cell lines EMT6/P and TA3 Hauschka would be models for the premalignant or early BC stage and an early or benign BC, respectively. With this cytogenomic information provided, these cell lines now can be applied really adequately in future research studies.


Introduction
Breast cancer (BC) is among the most common female specific cancer types and the second most common cancer in humans after lung cancer [1,2]. Survival rates of BC patients increased in the last years, especially in countries with early diagnostic regimens [3].
There are five stages of BC: (i) benign, premalignant stage; (ii) atypical ductal hyperplasia; (ii) preinvasive stage of ductal carcinoma in situ; (vi) metastatic carcinoma; and (v) advanced stage [4]. Known genetic changes in BC include acquired but also inherited changes in oncogenes, tumor suppressor genes and/or genes responsible for genomic stability [5,6]. Nowadays, such changes can be used as biomarkers for BC progression [7]. Moreover, BC can also be grouped according to immunohistochemical markers (ICM), like (i) presence or absence of receptors like those for estrogen (ER), progesterone (PR), human epidermal growth factor receptor-2 (HER-2) or epidermal growth factor receptor (EGFR) on tumor cell surface; (ii) expression of nuclear protein Ki67 as a marker of cell proliferation; and (iii) cytokeratin 5 expression in the plasma of BC cells. These and maybe more ICMs, Table 1. Relationship between the molecular breast cancer (BC) subtypes and immunohistochemical markers (ICMs) [11]. BC subtyping is extremely important for metastasis staging and treatment [7,10,12,13]. In the majority of BC cases, surgical excision of primary tumor is the initial treatment step. Afterwards, radio-and especially chemotherapeutic options are legion; accordingly, various "Clinical Decision Support Systems" are available based on which the potentially most advantageous treatment options for individual patients may be found [14]. Important to mention here is also that gene mutations and specific acquired molecular signatures recently had some impact on the advance of therapeutic targets in breast cancer treatment [15,16].

Molecular
However, the need for new types of medication with less side effects and being best targeted is still high [3]. Therefore, animal, especially murine, models are regarded as a highly feasible way not only to study biological pathways involved in initiation, progression and metastasis of a tumor like BC but also to establish new targeted medication, like murine tumor cell lines [16][17][18][19]. In spite of the widespread use of such murine tumor cell lines in research, surprisingly, most of them are not characterized molecular cytogenomically [20].
Fluorescence in situ hybridization (FISH) is considered the most practicable technique to detect gross genetic alteration in cancer [2]. Thus, in this study, it was taken advantage from combining multicolor-FISH using whole chromosome painting (wcp) probes, FISH banding [21], i.e., murine multicolor banding (mcb), and array-comparative genomic hybridization (aCGH) to do a first cytogenomic characterization of the three murine BC cell lines C-127I, EMT6/P and TA3 Hauschka, which is more than timely, as these cell lines were already established in 1978 [22], in 1986 [23] and in 1953 [24], respectively. Cell lines C-127I and EMT6/P were induced in mice by Harvey virus [22] and anthracycline treatment [23], respectively, while TA3 Hauschka was derived from tumorigenic ascites of a natural murine BC [24].
As previously done in comparable studies in murine tumor cell lines, a successful in silico translation from murine to human genome determined the corresponding homologous genetic alteration in human BC and enabled a classification as murine late stage, premalignant stage and benign BC-models.

C-127I
Murine BC cell line C-127I presented as a pentaploid but was basically genetically relatively instable and was rearranged with more than 10 derivative chromosomes. Consequently, this cell line can be divided to four clones, clone 1 being the ancestor clone and clones 2, 3 and 4 being derivatives of that.
For clone 3, representing 23% of the cells compared to clone 1, one chromosome 12 was lost and one was replaced by a derivative chromosome 12: del(12)(A1.1).
Clone 4 formed the remaining 17% of cells-here, an additional der(5)t(5;13) and a idic(18) were present as a structural aberrations and one chromosome 10 and 18 each were lost.

EMT6/P
This cell line EMT6/P is triploid and can be divided into five main clones which show some chromosomal instability. Clone 2 can possibly be considered the "ancestor" clone.
Clone 4 formed the remaining 17% of cells-here, an additional der(5)t(5;13) and a idic(18) were present as a structural aberrations and one chromosome 10 and 18 each were lost.

EMT6/P
This cell line EMT6/P is triploid and can be divided into five main clones which show some chromosomal instability. Clone 2 can possibly be considered the "ancestor" clone.

TA3 Hauschka
This cell line TA3 Hauschka had a particularly stable diploid karyotype and can be divided in three clones for which probably clone 1 was an ancestor of clones 2 and 3.

TA3 Hauschka
This cell line TA3 Hauschka had a particularly stable diploid karyotype and can be divided in three clones for which probably clone 1 was an ancestor of clones 2 and 3.
Kayrotype of clone 1 is shown in Figure 3 and can be described as follows:

aCGH Results
The aforementioned FISH studies of the three murine BC cell lines were in line with the aCGH results and are summarized in Figures 4-6a. A translation of those results to the human genome (only imbalances larger than 3.5 megabase pairs were included in the evaluation) identified the corresponding homologous region in the human genome (Figures 4-6b). Genomic details can be found in Supplementary Table S1. Clone 2 (30%) was characterized by an additional reciprocal translocation t(3;11)(G;E1) and a more complex derivative of chromosome 19, der(19)(19pter→19C2::XA6→XF5::8D1→8qter), compared to clone 1.

aCGH Results
The aforementioned FISH studies of the three murine BC cell lines were in line with the aCGH results and are summarized in Figures 4a, 5a and 6a. A translation of those results to the human genome (only imbalances larger than 3.5 megabase pairs were included in the evaluation) identified the corresponding homologous region in the human genome (Figures 4b, 5b and 6b). Genomic details can be found in Supplementary Table S1.

Comparison with Literature
The three studied BC cell lines present acquired copy number variations in regions known to harbor oncogenes and tumor suppressor genes, related to human BC [25,26]; as summarized in Table 2, gains of copy numbers were more frequent than losses. In Table 3, the chromosomal breakpoints observed in the three cell lines are compared to the break events known from human BC; here again, the highest rate of breaks being in concordance with human BC is present for the most advanced cell lines C-127I. Specific DNA copy number alterations correlated with the molecular subtype for human BC [27]-a comparison of the three murine BC cell lines is shown in Table 4. As a result, a high correspondence between C-127I and BC subtype HER2+ and basal-like tumors but also luminal B type was visible.

aCGH Results
The aforementioned FISH studies of the three murine BC cell lines were in line with the aCGH results and are summarized in Figures 4-6a. A translation of those results to the human genome (only imbalances larger than 3.5 megabase pairs were included in the evaluation) identified the corresponding homologous region in the human genome (Figures 4-6b). Genomic details can be found in Supplementary Table S1.

Comparison with Literature
The three studied BC cell lines present acquired copy number variations in regions known to harbor oncogenes and tumor suppressor genes, related to human BC [25,26]; as summarized in Table 2, gains of copy numbers were more frequent than losses. In Table 3, the chromosomal breakpoints observed in the three cell lines are compared to the break events known from human BC; here again, the highest rate of breaks being in concordance with human BC is present for the most advanced cell lines C-127I. Specific DNA copy number alterations correlated with the molecular subtype for human BC [27]-a comparison of the three murine BC cell lines is shown in Table 4. As a result, a high correspondence between C-127I and BC subtype HER2+ and basal-like tumors but also luminal B type was visible.  [25,26] and their involvement in gains or loss of copy numbers in the three studied cell lines. Abbreviation: CNV = copy number variant.

Comparison with Literature
The three studied BC cell lines present acquired copy number variations in regions known to harbor oncogenes and tumor suppressor genes, related to human BC [25,26]; as summarized in Table 2, gains of copy numbers were more frequent than losses. In Table 3, the chromosomal breakpoints observed in the three cell lines are compared to the break events known from human BC; here again, the highest rate of breaks being in concordance with human BC is present for the most advanced cell lines C-127I. Specific DNA copy number alterations correlated with the molecular subtype for human BC [27]-a comparison of the three murine BC cell lines is shown in Table 4. As a result, a high correspondence between C-127I and BC subtype HER2+ and basal-like tumors but also luminal B type was visible.  [25,26] and their involvement in gains or loss of copy numbers in the three studied cell lines. Abbreviation: CNV = copy number variant.    [25,26] and their involvement in gains or loss of copy numbers in the three studied cell lines. Abbreviation: CNV = copy number variant.  Table 3. Breakpoints in C-127I, EMT6/P and TA3 Hauschka compared to the observed acquired breaks in human BCs according to the literature [6,7,11,25,26,[28][29][30][31][32][33][34][35]: Concordances with human breakpoints are highlighted in bold.

Breakpoint Acc. to Human Genome
Human BC C-127I EMT6/P TA3 Hauschka

Discussion
BC has several subtypes ( Table 1) due to the heterogeneity and complex pathology [13]. Still, animal model systems play a major role in BC research, especially for testing new treatment protocols [4,13]. Thus, in this study, the three frequently used murine BC cell lines C-127I, EMT6/P and TA3 Hauschka were for the first time characterized on the molecular cytogenomic level. The feasibility of the applied scheme was shown in several previous studies [20,[36][37][38][39].
The three cell lines showed differences in the genomic alteration rate regarding ploidy, numerical and structural aberrations, and tumor-associated breakpoints. C-127I presented a very complex karyotype with pentaploidy, EMT6/P just was triploid and had less structural aberrations than C-127I, while TA3-Hauschka was near diploid and had only few imbalances. Thus, this is the first evidence that the three cell lines may demonstrate different subtypes and stages of human BC [40]. However, polyploidization is common in cancer cell lines [41,42], but polyploidization may also be part of malignancy progression and was also observed in connection with drug resistance [42]. Therefore, the pentaploidy in cell line C-127I may indicate that this cell line may be a suited model for the aggressive stage of BC or the HER2-enriched subtype which is known as an aggressive subtype and resistant to treatment [11,34].
As all three cell lines were established between 34 and 67 years ago [22][23][24], the chromosomal changes may be both original tumor related or acquired during long times of in vitro cultivation [43][44][45]. Only for TA3 Hauschka, the chromosome number at establishment is known, i.e., 41; thus, this cell line was, as other murine tumor cell lines, remarkably stable over times [24].
As highlighted in Table 2, CNVs could be observed in the three studied cell lines for 7, 10 or 17 of 21 regions known to be locations of tumor suppressor and/or oncogenes in human BC. The most aberrant was again C-127I, followed by EMT6/P and TA3 Hauschka. Interestingly, all three cell lines had gains of copy numbers in the region where the murine erbb2 gene is localized, while gain in the brca1 region was only evident for C-127I. Also, oncogene myc, being amplified in many human tumors, showed gain of copy numbers in C-127I and EMT6/P but not in TA3 Hauschka. Tumor suppressor gene rb1 was only deleted in C-127I.
As according to ploidy, karyotype and CNVs, C-127I is the most aberrant cell line, followed by EMT6/P and TA3 Hauschka; the same tendency is emphasized when looking at the number of chromosomal breakpoints and their localizations ( Table 3). The murine chromosomal band homologous to human 9p21, which harbors tumor suppressor genes important in retinoblastoma and p53 pathway regulation, is affected by chromosomal break events in all three studied murine cell lines. The aforementioned pathways are affected in aggressive forms of BC [28,46]. The murine homologous band to human 19p13.1 is involved also in break events in all three studied cell lines, and this region was identified to comprise genes correlated for enhanced BC risk [29,31,47]. Furthermore, mutations in this region are strongly associated only with ER-negative BC subtypes [33]. Moreover, in TA3 Hauschka and C-127I, breakpoints homologous to human band 14q32 could be observed. Notably, there, the gene DICER1 is located, recently identified as playing a role in cancer predisposition [30,32].
Horlings and coworkers correlated in 2010 [27] the frequency for gains and losses of CNVs to BC molecular subtypes. For this study, only C-127I could be clearly assigned and not only to one but to even three BC subtypes (Table 4): HER2+, basal-like tumors, and luminal B type.
Finally, one may speculate about the influence of breakpoints and/or CNVs of regions being involved in presenting ICM, like ER or PR; in none of the three cell lines, there are changes in regions being homologous to gene estrogen receptor 1 (ESR1; in human 6q25.1). However, there was a loss in gene estrogen receptor 2 (ESR2; in human 14q23.2~23.3) in C-127I. The latter cell line presents also gains in progesterone receptor gene (PR; in human 11q22.1).
In conclusion, the molecular cytogenomic study and in silico translation for the three here studied BC cell lines, C-127I, EMT6/P and TA3 Hauschka, revealed that they can be used as models for human BC at different stages of malignancy. TA3 Hauschka can be best considered as a model for benign BC, EMT6/P may be used to represent the premalignant or early malignant stage of BC, while C-127I can be the model for the advanced BC stage. These insights are important for future application of these cell lines in BC research and their adequate use.

Cell Lines
The fibroblast-like/adhesively growing cell lines C-127I (Catalogue number CLS 400134) was purchased from CLS Cell Lines Service GmbH (Eppelheim, Germany), while the EMT6/P-(Catalogue number ECACC 96042344) and TA3 Hauschka-cell line (Catalogue number ECACC 85061102) were obtained from European Collection of Authenticated Cell Cultures (Salisbury, UK). They were grown adherently in DMEM medium (C-127I) or in EMEM/EBSS medium (EMT6/P and TA3 Hauschka) according to company instructions. Cells were prepared cytogenetically as previously reported [20]; whole genomic DNA was extracted using the Blood & Cell Culture DNA Midi Kit (Qiagen, Hilden, Germany) [36]. Cell line-derived chromosome-preparations were subjected to molecular cytogenetic analysis and extracted DNA were subjected to aCGH analysis. Cells were harvested shortly before they reached confluency; other cell density tests were not undertaken.

Data Analysis
The breakpoints and imbalances of the three studied BC cell lines were determined according to aCGH and mcb data and aligned to human homologous regions using Ensembl Browser, as previously described [37]. The obtained data was compared to genetic changes known from human BC according to literature mentioned.

Acknowledgments:
The technical support of Nadezda Kosyakova (Jena, Germany) is kindly acknowledged.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.