Human Breast Cancer Cell Lines Co-Express Neuronal, Epithelial, and Melanocytic Differentiation Markers In Vitro and In Vivo

Differentiation programs are aberrant in cancer cells allowing them to express differentiation markers in addition to their tissue of origin. In the present study, we demonstrate the multi-lineage differentiation potential of breast cancer cell lines to express multiple neuronal/glial lineage-specific markers as well as mammary epithelial and melanocytic-specific markers. Multilineage expression was detected in luminal (MCF-7 and SKBR3) and basal (MDA-MB-231) types of human breast cancer cell lines. We also observed comparable co-expression of these three cell lineage markers in MDA-MB-435 cells in vitro, in MDA-MB-435 primary tumors derived from parental and single cell clones and in lung metastases in vivo. Furthermore, ectoderm multi-lineage transdifferentiation was also found in human melanoma (Ul-MeL) and glioblastoma cell lines (U87 and D54). These observations indicate that aberrant multi-lineage transdifferentiation or lineage infidelity may be a wide spread phenomenon in cancer.


Introduction
Cancer diagnosis of the tissue origin of metastatic lesions, especially those from occult primary tumors, relies heavily on the expression of cellular or tissue differentiation markers. Emerging clinical and pre-clinical evidence show that the differentiation programs are aberrant in cancer cells allowing them to express differentiation markers beyond their tissue of origin [1][2][3]. These observations have implications for both cancer research and clinical management of cancer. However, this property has not been thoroughly examined in human cancer cell lines that are frequently used in cancer research.
A microarray analysis has indicated that the gene expression pattern of the human MDA-MB-435 [4] resembles that of human melanoma cell lines [5]. Since then, additional evidence has shown the ability of MDA-MB-435 cells to express melanocytic markers [6][7][8]. However, upon induction with heregulin in vitro, MDA-MB-435 cells undergo sufficient mammary epithelial differentiation to produce milk lipid droplets and to express b-casein mRNA [9]. Additionally, enhanced expression of NM23-H1 metastasis suppressor gene leads to the formation of organized mammary acinus-like sphere in 3D culture and the expression of sialomucin (epithelial membrane antigen, EMA) [10]. A recent investigation confirmed the ability of MDA-MB-435 cells to co-express markers of mammary epithelium and melanocytes both in vitro and in vivo, and postulated committing lineage infidelity as an underlying mechanism for the observed dual lineage transdifferentiation [11]. Furthermore, another recent study reported the robust expression of melanocyterelated genes in a variety of breast cancer cell lines including MDA-MB-435, and more importantly in freshly resected and histopathologically confirmed human breast cancer specimens [12].
Here, we report the expression of multiple neuronal/glial lineage-specific markers by MDA-MB-435 cultured cells in vitro and in primary tumors and lung metastases in vivo in addition to the reported expression of epithelial and melanocytic markers. Furthermore, we observed the co-expression of three ectoderm cell lineage markers in other luminal (MCF-7 and SKBR3) and basal (MDA-MB-231) types of human breast cancer cell lines, as well as in human melanoma and glioblastoma cancer cell lines generated from tumors of ectoderm origin.
These observations indicate that while terminal differentiation to the anticipated cellular type is compromised in the cancerous state, aberrant multi-lineage transdifferentiation or lineage infidelity may be a wide spread in cancer phenomenon.

Cell culture
The MDA-MB-435s, MDA-MB-231, MCF-7 and MCF-10A cell lines were obtained from ATCC (Manassa, VA). SKBR3 cells were gifts from Dr. Suzanne Conzen, University of Chicago. Ul-Mel, U87, and D54 were from gifts from Dr. Ralph Weichselbaum, University of Chicago. All cells were cultured in DMEM high glucose (Hyclon, Logan, Utah) supplemented with 10% FBS and 1% penicillin/streptomycin. MDA-MB-435-GFP derivative cell lines were generated from single cell cloning through initial seeding of single cells in 96-well plates and subsequent expansion of cell numbers. MDA-MB-435-GFP-L cell lines were generated from resected lungs that harbor pleural metastases derived from tail-vein injected MDA-MB-435-GFP cells and purified using G418 for the selection of GFP expression.

PCR analysis of lineage markers expression
Total RNA was isolated using TRIZOL (Applied Biosystems Inc. Foster City, CA) according to the manufacturer's instructions, followed by DNAse treatment (Promega, Madison, WI). cDNA was reverse transcribed from 2 mg of total RNA using random primer method (Applied Biosystems Inc.) according to the manufacturer's instructions. PCR amplification was performed by incubation at 94uC for 1 min, followed by 32 cycles of 94uC for 30 s, 55uC for 30s, and 72uC for 25s. Amplified products were separated and visualized on a 0.8% agarose gel. TBP was used as a loading control.

Immunohistochemistry (IHC) analysis of differentiation marker expression in cultured cells, primary tumors, and lung metastases
All animal studies were carried out according to protocols approved by the IACUC Committee at the University of Chicago. Six-week-old female athymic Ncr/nu/nu mice (NCI-Frederick), 18 to 20 g, were used. For orthotopic tumor implantation, 5610 6 cultured MDA-MB-435 cells suspended in 0.1 ml of PBS were injected into the left inguinal mammary fat pad (m.f.p.). Tumors were harvested upon reaching ,200 to 250 mm 3 , fixed in formalin and embedded in paraffin. To produce experimental lung metastases, 1610 6 cultured MDA-MB-435-GFP cells, suspended in 0.1 ml of PBS, were injected into the mouse tail vein. Twelve weeks later, mouse lungs were dissected and fixed in formalin and embedded in paraffin. The tumor and lung blocks were cut into 5-mm paraffin sections. For IHC analysis, tissue sections were deparafinized, rehydrated followed by antigen retrieval and treated with 3% hydrogen peroxide to block endogenous peroxidase activity. 1:100 dilution of cytokeratin (DAKO, M3515), 1:100 dilution of melan-A (DAKO M7196), 1:100 GFAP (R & D System), and 1:500 dilution of Nestin (Abcam, ab5968) were applied to the tissue slides and incubated for one hour. Envision+anti-mouse system was used for antigenantibody detection. The slides were counter-stained with hematoxylin, air dried, and examined under light microscopy.  and 5]. The lack of expression of both CK17 and CK19 epithelial differentiation markers in cultured and non-induced MDA-MB-435 cells is consistent with its characterization as a poorly differentiated cancer cell line [4]. These results demonstrate that the expression of luminal and basal epithelial differentiation is aberrant in human breast cancer cell lines.

Results and Discussion
Thereafter, we determined the expression of three melanocytic differentiation markers (MITF, melan-A and tyrosinase) in breast cancer cell lines. We included human melanoma cell line Ul-MeL as a positive control for this analysis (Figure 2, lanes 6).   Figure 3B). Compared to the Ul-MeL melanoma cell line that expressed high levels of melan-A ( Figure 3A), the protein expression level of this melanocytic differentiation marker was considerably lower in MDA-MD-435 and U87 cells ( Figure 3A). In contrast, cultured MDA-MB-435 cells strongly expressed both GFAP and nestin proteins. Similar patterns of IHC staining of GFAP and nestin were observed (Figure 3). The discrepancy between mRNA and protein expression of cytokeratin and GFAP could be due to differences in mRNA stability among genes analyzed. This observation highlights the importance of using multiple complementary analyses for drawing conclusions from gene expression assessment.
Further, we also observed aberrant transdifferentiation or lineage infidelity in the Ul-MeL melanoma cell line that coexpresses epithelial and neuronal markers [ Figure 2(I) (III), lanes 6], and in two glioblastoma tumor cell lines that co-express epithelial and melanocytic markers [ Figure 2(I) (II), lanes [7][8].
These findings indicate that terminal differentiation to the anticipated cellular type is altered in the cancerous state and that the phenomena of lineage infidelity that is associated with the ability of cancer cells to transdifferentiate, occurs in different cancer types and is not limited to breast cancer.

Co-expression of multi-lineage protein markers in MDA-MB-435 primary tumors and metastases
To further confirm that the expression of epithelial, melanocytic and neuronal/glial markers in MDA-MB-435 cells is not due to an artifact of cell culture, MDA-MB-435 primary tumors growing orthotopically in the mammary fat pad (m.f.p) of nude mice and pleural macro-and micro-metastases of MDA-MB-435 were utilized for the analyses of the expression of three cellular lineage markers by IHC (Figure 4 and Methods). We consistently observed more robust and consistent MDA-MB-435 tumor growth (higher tumor cell take rate, initiation of tumor growth, higher level of tumor angiogenesis and higher incidence of lung metastasis) at the m.f.p, than when MDA-MD-435 was transplanted under the skin of the hind leg or on the back. These observations demonstrate that the m.f.p represents a more favorable microenvironment for MDA-MB-435 tumor initiation, expansion and progression, consistent with its ability to undergo mammary epithelial differentiation [9,14]. While cultured MDA-MB-435 cells were negative for AE1/AE3 staining for pan-  Figure 4A) and strong staining for ESA. This observation is consistent with a previous report [11]. Additionally, we also observed evidence of mammary duct differentiation at the periphery of MDA-MB-435 tumors ( Figure 4A, H&E panel). Similar to IFC, IHC and PCR analyses in vitro, both the orthotopic primary tumor and macro/micro lung metastases displayed intense and homogenous expression of nestin and GFAP ( Figure 4). In contrast, strong melan-A staining was only detected in discrete clusters of cancer cells within the primary tumor ( Figure 4A).
Based on the close clustering of MDA-MB-435 cells with melanoma cell lines in microarray analysis [5], the possibility of the breast cancer patient, from whom MDA-MB-435 was derived from, having an undetected occult melanoma was suggested. However, cell lines generated from breast tumors and from nonsmall lung carcinoma used in the same micro-array study also failed to yield a clear clustering pattern according to their tissue of origin. In particular, two other invasive breast cancer cell lines, Hs578T and BT-549 were clustered together with brain tumor cell lines, a finding that is consistent with our detection of neuronal and glial differentiation markers in breast cancer cell lines. Further, a recent report demonstrated a wide spectrum of expression of melanocyte-related genes in histologically confirmed human breast tumor specimens [3]. It thus appears that aberrant co-expression of multi-lineage markers via transdifferentiation or lineage infidelity occurs frequently in breast cancer. Therefore, molecular signatures derived from gene expression profiling should not be used as exclusive evidence or criteria to determine the tissue origin of a cancer cell line or a metastatic lesion from an occult primary tumor. Functional properties such as in vitro functional characterization of cellular differentiation (acinus formation in 3D gel culture, production of milk products upon induction of lineagespecific differentiation), in vivo tumor growth and progression should all be taken into account when we consider the classification of a human cancer cell line.
We have demonstrated in this report the co-expression of three ectoderm cell lineage differentiation markers by a panel of breast cancer cell lines, by MDA-MB-435 cells obtained from the ATCC, by cell lines derived from single-cell cloning of MDA-MB-435 and MDA-MB-435 cells derived from lung metastases grown in vitro; and as well as by MDA-MB-435 orthotopic primary tumors and lung metastases in vivo. It is thus highly unlikely that this cell line was contaminated by both melanoma cells and neuron/glial cancer cells, or that the breast cancer patient from whom MDA-MB-435 cells were derived also had an undiagnosed melanoma and undiagnosed glioblastoma.
In conclusion, our observations indicate that aberrant multilineage transdifferentiation or lineage infidelity occurs frequently in multiple types of human cancer and may be a wide spread phenomenon.