The successful primary culture of cancer cells from breast tumours has great significance for the creation of an original cell source to study breast cancer cells (BCC) biology and for the development of therapeutic strategies. Although many studies on cancer cell biology have been conducted using cancer cell lines Gillet et al., 2011Gillet et al., 2013Lacroix and Leclercq, 2004Neve et al., 2006, recent reports have highlighted that those cancer cell lines do not, for various reasons, consistently display original cell characteristics Keller et al., 2010. Therefore, to ensure that a chosen cell culture model accurately reflects the biological characteristics of the cells of interest, the use of primary cancer cells is essential.

Numerous primary culture studies are underway to identify optimal conditions, with regard to efficiency and duration, for the acquisition of BCCs. While there are many factors that influence the primary culture process, the culture medium is the key parameter. In studies on HBC cells, culture conditions were initially adopted from a number of similar studies on normal HME cells that had identified important factors determining their in vitro survival Hammond et al., 1984Stampfer et al., 1981aStampfer et al., 1981b. Regrettably, medium components for the propagation of primary human breast cells (HBC) have not yet been determined, as the biology of this cell type remains largely unclear. Stampfer et al. Stampfer, 1982Stampfer and Bartley, 1985Stampfer et al., 1993 developed a variety of culture media for the growth of human mammary epithelial (HME) cells; the original medium consisted of several undefined components, but was later refined to a hormone- and growth factorsupplemented medium that supports proliferation of HME cells over many in vitro passages. Band and Sager Band and Sager, 1989 showed that it was useful to propagate HME cells extensively in a growth factor- and hormone-supplemented medium that also contained serum and pituitary extract. Subsequently, Petersen and Van Deurs Petersen and van Deurs, 1987 and Ethier et al. Ethier et al., 1991 reported the growth of normal HME cells in serum-free media in the absence of pituitary extract or serum. Although these culture media promoted the growth of normal mammary epithelial cells, very few of them supported the growth of BCCs Bartek et al.,1985Taylor- Papadimitriou et al., 1989. Thus, the culture conditions that are conducive to the rapid proliferation of normal HME cells over many in vitro passages hardly support the growth of BCCs Wolman et al., 1985.

The use of a serum-free medium for the cultivation of normal HME cells circumvents a number of problems that are mainly related to the instability of serum Hammond et al.,1984Smith et al., 1981. However, in comparison with serum-containing medium, serum-free medium also entails some disadvantages such as the need for a complex mixture of highly pure medium components and a reduced cell proliferation rate. Mammary tumour cell lines have been isolated and grown in standard medium (e.g. Dulbecco’s Modified Eagle Medium/Nutrient Mixture F-12, DMEM/F12) supplemented with 10% foetal bovine serum (FBS) Engel and Young, 1978Smith et al., 1987Soule et al., 1973. Serum contains growth factors, which promote cell proliferation, as well as adhesion factors and antitrypsin activity, which promote cell attachment. However, it has recently been accepted that tumours consist of highly heterogeneous cell populations with respect to cellular morphology, proliferative potential, genetic lesions, and treatment response Bomken et al., 2010Marusyk and Polyak, 2010. The isolation of cells from breast tumours may give rise to several different cell types; normal counterparts from which the neoplastic cells arise, such as connective-tissue fibroblasts, infiltrating immune cells, vascular endothelial cells, and smooth muscle cells, as well as other elements of the normal tissue can all survive explantation Sung et al., 2007Weber and Kuo, 2012Yu et al., 2011. Therefore, breast cancer epithelial cells in tumour biopsies are irrevocably overgrown by fibroblasts in a medium supplemented with high serum concentrations.

In any case, the success rate of a BCC culture is increased greatly by using selective media that enrich the population of BCCs, but prevent the rapid and extensive growth of normal cells, including stromal cells and normal HME cells. Ethier et al. found that the addition of 5% FBS to medium supplemented with insulin, hydrocortisone, EGF, cholera toxin, and progesterone stimulated rapid proliferation of breast cancer epithelial-like cells Ethier et al., 1993. They also found that a relatively simple medium, only supplemented with 5% FBS, insulin, and hydrocortisone, resulted in the slow emergence of BCCs that ultimately gave rise to BCC lines Ethier et al., 1993.

To date, the essential characteristics of primary breast cancer cells are still a matter of debate. In particular, recent studies have shown that solid breast tumours harbour a cell population with stem cell characteristics, which is responsible for the formation and maintenance of tumours, development of metastases, and, eventually, patient mortality. These cells are known as cancer stem cells Al-Hajj et al., 2003Clarke, 2005. In this study, we aimed to develop a standardized protocol for the isolation and propagation of HBC cells from primary tumour biopsies that a) ensures that the isolated primary cells include breast cancer stem cells and b) minimizes contamination with other cells.

Breast tumours contain a combination of various kinds of cells, including normal epithelial cells, stromal cells, breast cancer cells, and breast cancer stem cells. A suitable protocol for the isolation of BCCs must not only provide for high cell growth efficiency, but also for the establishment of cells exhibiting BCC properties. From the information gathered from previous studies investigating single cell culture versus explant tissue culture, this study employed expanding tissue culture (data not shown). In this method, the culture medium is the most decisive factor in the outgrowth of cells from tumour fragments, as well as in the types of cells obtained. Based on existing literature reports, 3 different kinds of media were chosen for use in this study. M171 medium supplemented with MEGS (Medium M) is a serum-free medium that supports the proliferation of normal human epithelial mammary cells. In contrast, DMEM/F12 medium supplemented with 10% FBS (Medium D) is a serumcontaining medium that supports the proliferation of routine human BCC lines. For the purpose of this study, these 2 kinds of medium were mixed in a ratio of 1:1 to produce a third medium (Medium DB). Thus, Medium DB contained 50% of each of the components of Medium D and Medium M, and the serum concentration was similarly reduced to 5%.

As detailed in the Results section, Medium D was not suitable for the isolation of BCCs. Although in this medium the cells migrated more rapidly than in Medium M and Medium DB, they also showed a relatively slow mitosis rate and therefore reduced proliferation. Hence, in nearly all samples, cells grown in Medium D were not sufficiently high in number for use in further evaluation. In contrast to Medium D, primary cells cultured in Medium M proliferated rapidly, and the majority of samples cultured in Medium M provided enough cells for additional experiments. This observation can be explained by the fact that Medium M contained a pool of hormones and growth factors such as hydrocortisone, EGF, and insulin. These factors are beneficial for the survival and proliferation of breast tissue-derived cells. Primary cells cultured in Medium DB proliferated most rapidly, such that all samples provided cells that were sufficiently high in number for use in further experiments. These results are likely due to the components of Medium DB, which included growth factors and hormones from Medium M, as well as serum from Medium D. However, it should be highlighted that the serum concentration is reduced in comparison to common serum levels, and that this appears to be beneficial with regard to the elimination of stromal cells. Therefore, using Medium DB for primary tumour cell culture results in the rapid proliferation of primary cell populations.

Next, the existence of a CD44⁺CD24^- population was evaluated by flow cytometry in all primary cell cultures. Nearly all primary cells grown in any of the 3 culture media tested positive for CD44 and negative for CD24, with the highest percentage of CD44⁺CD24^-cells occurring in Medium D and Medium DB. In a previous study by Al-Hajj et al. (2003), primary cells contained a sub-population of CD44⁺CD24^- cells with high tumourigenicity Al-Hajj et al., 2003. However, the possibility that the primary cells expressing CD44⁺CD24^- were not all BCSCs appears likely Ghebeh et al., 2013Mannello, 2013. In a recent study, Ghebeh et al. (2013) clearly demonstrated that both normal breast tissue and breast cancer tissue harboured CD44⁺CD24^- cells Ghebeh et al., 2013. They also suggested that BCSCs would be enriched in the CD44⁺CD24^-cells if they were combined with the CD49f⁺phenotype Ghebeh et al., 2013. CD49f was also determined as a marker of BCSCs in previous studies Meyer et al., 2010Yu et al., 2012. Therefore, in the next experiment, the existence of a CD49f⁺cell population in the primary cells was evaluated. The results showed that primary cultures in Medium M contained the highest percentage of CD49f⁺cells, followed by primary cultures in Medium DB and Medium D. Hence, Medium M efficiently supported the growth of cells with the CD49f⁺phenotype; however, compared with Medium D and Medium DB, Medium M did not support the growth of CD44⁺CD24^- cells. Medium DB excellently promoted the growth of CD44⁺CD24^- cells, but little impact on the proliferation of CD49f⁺cells.

However, Medium D and Medium M also supported stromal cell proliferation. Regarding CD90 expression, more than 50% of the primary cells in Medium D and Medium M tested positive for this marker, whereas this population only accounted for about 25% of the primary cells grown in Medium DB. These CD90⁺cells were considered as contaminant cells in the breast carcinoma primary culture Araki et al., 2007Haack- Sorensen et al., 2008Nakamura et al., 2006. In sum, a marked contrast was observed with respect to the proportion of contaminant cells versus cells with the BCSC phenotype (CD44⁺CD24^- and CD49f⁺) in primary cultures grown in Medium DB and Medium M.

Next, cellular ALDH expression was monitored to evaluate culture efficiency. The ALDH enzyme has important functions in the development of epithelial homeostasis, and deregulation of this class of enzymes has been implicated in multiple cancers Marchitti et al., 2008. The ALDEFLUOR assay is thought to be an almost universal marker of stem cell activity in both normal and cancer tissues Corti et al., 2006Hess et al.,2004, including normal and malignant breast epithelial stem cells Ginestier et al.,2007. In this study, the ALDH⁺cell population was approximately 5 times larger in Medium DB than in Medium M (9.35% ± 3.64% and 2.28% ± 0.88%, respectively). To sum up, Medium DB, significantly more than the other 2 media, specifically promotes the growth of the breast cancer stem cells that exist in malignant breast tumours. To support this conclusion, karyotype analysis revealed that nearly all cells in Medium DB exhibited an abnormal karyotype, while in Medium M as well as Medium D, primary cells contained both normal and slightly abnormal karyotypes. All samples doing karyotype derived from female breast cancer patients whose tumors were diagnosed as primary tumors, i.e. they had never undergone any previous treatment, including chemotherapy or radiotherapy. Consequently, the number of chromosomes of primary cancer cells was not so different than the normal chromosome number, known as 46 chromosomes. This is consistent with many studies of cancer cells primary culture, including Adeyinka et al., 2000Bardi et al., 1993Brothman et al., 1990Ferti et al.,2004Stamouli et al., 2004Teixeira et al., 1995.

Following karyotyping, primary cells from Medium M and Medium DB were used to induce tumours in mice. The results showed that primary cells grown in either Medium M or Medium DB, as well as other BCC lines such as MCF-7 and MDA-MD-231 successfully caused tumours in mice when injected at a cell density of 10⁶ cells per mouse. Lower densities of primary cells or BCC lines failed to establish tumours in mice. Mouse models that were used in the experiments of examining the dose causing tumors were athymic nude mice, whose immune system is partially suppressed. Therefore, human cell transplantation, primary cells or BCC lines, induced immune response in mice, with the most powerful after 1 week. As a result, grafted cells including primary cells and BCC lines existed only 2 weeks. However, the results also showed the ability to establish tumors in mouse model with primary cells cultured in Medium DB and Medium M was as well as BCC lines. Tumour sections stained with HE confirmed that the tumours contained cancer cells with large nuclei; the tumours were established from primary cells cultured in Medium DB and Medium M.

Moreover, after long-term cultivation of approximately 6 months in Medium DB, breast cancer cells were observed to undergo the process of MET, i.e. altering their mesenchymal phenotype back to the epithelial phenotype. This process continued until epithelial cell colonies were formed, and eventually all cells in the culture flask transformed into the epithelial phenotype. At this point, cells proliferated more rapidly and stably, in a manner similar to that observed in typical BCC lines. Epithelial cancer cells acquire mesenchymal features that provide a mechanism for tumour cells to leave the primary tumour, resulting in the induction of single cell and/or collective cell migration Drasin et al., 2011Mani et al., 2008May et al., 2011. Overall, the carcinoma epithelial-mesenchymal transition (EMT) is defined by a loss of normal epithelial architecture, which renders the epithelial tumour cells phenotypically indistinguishable from fibroblasts. At the molecular level, this is characterized by a downregulation of epithelial markers and an up-regulation of mesenchymal markers, accompanied by an increase in cell migration and invasion Polyak and Weinberg,2009Scheel and Weinberg, 2011Thiery et al., 2009. However, the role of mesenchymal-epithelial transition (EMT) in cancer is complicated by the fact that in the appropriate microenvironment, mesenchymal-like cells likely undergo a reversion or MET, thereby permitting colonization Micalizzi et al., 2010Wells et al., 2008.

BCCs are essential biological tools for both research and therapy. Breast tumours always contain a combination of different kinds of cells, including breast cancer cells. This study successfully established a simple cell isolation procedure based on breast tumourexplant cultivation in Medium DB, which is a 1:1 mixture of DMEM/F12 supplemented with 10% FBS and M171 supplemented with 1X MEGS. Cells isolated using this procedure exhibited BCC properties such as expression of BCSC markers (CD44⁺CD24^-), expression of ALDH, abnormal karyotype, and tumourigenic capability in mouse models. The findings of this study served to establish a method that proved highly useful for the isolation of BCCs from tumour biopsies as well as for the enrichment of BCSCs.

ALDH: Aldehyde dehydrogenase, APC: Allophycocyanin, BCC: Breast cancer cell, BCSC: Breast cancer stem cell, BPE: Bovine pituitary extract, DEAB: Diethylaminobenzaldehyde, DMEM/F-12: Dulbecco’s Modified Eagle Medium/Nutrient Mixture F-12, EGF: Epithelial growth factor, EMT: Epithelialmesenchymal transition, FBS: Fetal Bovine Serum , FCM: Flow cytometry, FITC: Fluorescein Isothiocyanate, HE: Hematoxylineosin, HBC: Human breast cancer cell, HME: Human mammary epithelial cells, Medium D: DMEM/F12 supplemented with 10%FBS and 1% Antibiotics/antimycotic, Medium M: Medium 171 supplemented with MEGS and 1% Antibiotics/antimycotic, Medium DB Medium D and Medium M (1:1, v/v), MEGS: Mammary Epithelial Growth Supplement, MET: Mesenchymal-epithelial transition, NOD/SCID: Non-obese diabetic/severe combined immune deficiency.