Background
ALCAM/CD166 is an immunoglobulin cell adhesion molecule expressed by neuronal, endothelial, hematopoietic and epithelial cells [
1‐
13]. It's up-regulation in cancer was first identified at the RNA level in melanoma cell lines as
memD [
14]. Subsequently, increased ALCAM expression was found in melanoma tumors
in situ [
13,
15]. More widespread deregulation of ALCAM expression has since been reported in several other tumors including those of the prostate [
16,
17], esophagus [
18], colon [
19], bladder [
20] and pancreas [
21]. Alterations in ALCAM expression in tumors have recently been reviewed by Ofori-Acquah and King [
22].
In a study of primary breast cancer tissues and non-neoplastic mammary tissue from the same mastectomies, we discovered that the level of ALCAM transcripts was lower in breast cancer tissues from patients who had metastases to regional lymph nodes [
23], and that primary tumors from patients who died of breast cancer had significantly lower levels of ALCAM transcripts [
23]. Subsequent studies showed that patients with the lowest level of ALCAM transcripts develop skeletal metastasis [
24], that low ALCAM correlated with an aggressive tumor phenotype and significantly negative correlation between ALCAM expression and tumor diameter and grade [
25]. More recently high-level ALCAM in breast cancer tissues has emerged as a predictor of good outcome among patients treated with tamoxifen [
26] and adjuvant chemotherapy [
27,
28].
Tumor cells circulate in blood as single entities and multi-cellular emboli [
29], and form secondary colonies in the vascular wall. This mechanism of metastasis is supported by evidence showing that tumor cells perfused in isolated rat lungs attach to the endothelia wall with minimum extravasation, leaving the endothelium-attached cells as the seeds of secondary tumors [
30]. Indeed, in primary tumors derived from subcutaneous injection of murine breast carcinoma cells in immunocompromised mice, early metastatic colonies are intravascular in origin [
31]. That adhesion molecules tethered on tumor cell surfaces influence their colonization of the lung, and downstream metastatic processes, is supported by the finding that the loss of ALCAM at the cell surface confers a high risk for disease progression and mortality in nodal negative cases of breast cancer [
26].
In this study, the ALCAM gene was cloned and functionally characterized in a panel of breast cancer and melanoma tumor cell lines, and the influence of ALCAM on homotypic tumor cell adhesion in the pulmonary vasculature investigated. Our findings provide new mechanistic insights on ALCAM that can be developed further to alter its negative influence in tumor cell progression.
Discussion
Altered expression of the cell adhesion molecule ALCAM is associated with progression, metastasis and response to therapy in multiple cancers, yet specific DNA elements that regulate ALCAM promoter activity have not previously been defined. We found a consensus NF-κB element at -1140 and site-directed mutagenesis of this element significantly reduced ALCAM promoter activity. Over-expression of p65 NF-κB increased ALCAM promoter activity, while p65 occupied the cognate motif on the endogenous ALCAM promoter in melanoma cells lines (Fig.
4A ii). These data strongly suggests that ALCAM is a target of the NF-κB pathway. Our data in melanoma cells is consistent with previous reports showing that transformation of avian lymphoma B-cells with
v-rel induces ectopic expression of ALCAM [
32]. Mutations of cell cycle genes activate NF-κB, which in turn is directly responsible for increasing expression of several target genes involved in melanogensis [
33‐
36]. Given the marked elevation of expression of ALCAM in vertically growing primary melanoma tumors, it is likely that the ALCAM gene locus is a downstream target of NF-κB in melanoma. Our reporter gene analysis did not reveal
cis-active elements in the ALCAM promoter that suppress ALCAM expression in breast cancer cell lines. However, the human ALCAM gene locus is relatively large [
37] and may contain negative regulatory elements far upstream of the transcription start site, where they may influence gene expression, via long-range DNA looping.
Epigenetic modification of DNA is a well established mechanism for suppressing genes, and we discovered that the proximal ALCAM promoter is endowed with multiple CpG islands, which are targets for DNA methylation (Fig.
5A). Sequencing revealed that virtually all CpG islands in the ALCAM promoter are methylated in tumor cells lacking ALCAM expression. The plasticity of this modification, and its inherent linkage to gene expression was confirmed in experiments with 5-aza-deoxycytidine (Fig.
6A, B). Additional correlative data in both breast cancer and melanoma cell lines showed direct relationship between the level of DNA methylation of the ALCAM promoter and ALCAM protein level (Fig.
5B). 5-Aza-deoxycytidine is currently in clinical use, and may be indicated to boost ALCAM expression in breast cancer [
38], however, it also alters expression of a large number of genes [
39,
40]. Moreover, our data suggests this drug alone maybe insufficient to fully restore ALCAM expression in tumors with severely repressed ALCAM expression. The level of ALCAM expression is variable in different tumor types. This heterogeneity can be resolved by the unique tissue origins of different tumors, although variable expressions have been described at different stages of tumor development in the same type of malignancies. With respect to breast cancer, multiple studies have examined ALCAM expression at the transcript and protein levels, using a variety of methods. There is an emerging consensus that low level ALCAM is a bad prognostic marker in breast cancer [
23,
24,
27,
41‐
44]. In the largest, and most recent study, Ihnen
et al found that low ALCAM mRNA was associated with shorter disease free survival and duration of survival in hierarchical cluster analysis involving training and multiple validation cohorts of breast cancer patients [
44]. This emerging paradigm is supported by the correlations of high ALCAM mRNA with progesterone and estrogen receptor status, better response and longer overall survival. It is reasonable to presume that down-regulation of ALCAM activates alternative compensatory pathways. Plausible candidates are other ALCAM isoforms, such as soluble ALCAM, which may account for the elevation of serum ALCAM in patients with breast cancer [
45]. A similar mechanism may explain the marked increased in cytoplasmic ALCAM in some aggressive breast cancer tissues [
26]. While the relationships between these various deregulations of ALCAM expression remain to be verified, they all represent loss of function of ALCAM, which to date has consistently been associated with poor prognosis in breast cancer.
Alteration of adhesion molecule expression is a hallmark of several cancers [
46], however the underlying biological mechanism for the deleterious effects of reduced ALCAM in breast cancer is poorly defined. In this study, we examined the impact of ALCAM on the adhesive behavior of tumor cells in the pulmonary vasculature using the isolated rat lung system. Experiments using two function blocking ALCAM antibodies, and genetically-modified MDA-MB-435 clones ectopically expressing ALCAM, revealed that ALCAM promotes homotypic tumor cell adhesion demonstrated by clusters of tumor cells in the pulmonary vasculature. ALCAM is localized at sites of lateral cell contacts in the pulmonary endothelium, and is therefore unlikely to mediate adhesions between circulating tumor cells and the endothelium [
47]. Colonization of the lung by tumor cells involves their interaction with the endothelial wall, and subsequent extravasations. The data reported here suggests that ALCAM is likely to slow down this process, by promoting homotypic tumor cell adhesion. This idea is pertinent to the level of ALCAM expression and the metastatic phenotypes of the two tumor cells lines we used in our lung perfusion experiments. The ALCAM-positive MDA-MB-231 cells which formed large cell clusters in the rat lung cannot metastasize to distant sites when injected into the mammary fat pad of athymic nude mice [
48,
49]. On the contrary, ALCAM-negative MDA-MB-435 efficiently and spontaneously forms distant metastasis under identical experimental conditions [
48‐
51]. Additional mechanistic studies are needed to clearly define the relationship between ALCAM level and the metastatic phenotypes of tumor cells. In this study, we have shown that loss of ALCAM function in MDA-MB-231 and gain of ALCAM function in MDA-MB-435, switches their adhesive phenotypes in the pulmonary vasculature, a process that influences metastasis to the lung [
30,
31].
Methods
Cells
Cells studied included MDA-MB-435, originally classified as of breast cancer origin, but recently shown to posses melanocytic lineage, fifteen breast cancer (BT549, BT483, MDA-MB-231, HCC70, HCC1428, HCC1806, MDA-MB-453, MCF-7, CAMA-1, MB-157, MDA-MB-361, HCC1500, HCC1008, T47D and SK-BR-3) and two normal epithelial breast cell lines (MCF-10A and MCF-12A) purchased from American Type Culture Collection (ATCC, Rockville, MD). In addition, four melanoma cell lines (FEMX-I, LOX, MelJuso, C8161.9) were studied. All cell lines were cultured using media conditions recommended by the commercial or academic suppliers. Cells were incubated at 37°C in a humidified chamber with 5% CO2 (except for MDA-MB-361, which was cultured in room air). MDA-MB-435 cells (2 × 105 per 35-mm well) were plated overnight, and then treated with a concentration range (0-10 μM) of 5-aza-2-deoxycytidine (EMD, Madison, WI). Cells were replenished with fresh medium containing 5-aza-2-deoxycytidine every 48 hours for six days, and harvested for analysis.
Western Blots
Cell lysates (20 μg) were resolved on a 10% SDS-PAGE. Proteins were transferred to PVDF membrane and probed with antibodies (ALCAM/CD166, Novocastra Laboratories) and secondary antibody conjugated to horseradish peroxidase and detected using chemiluminescence (Pierce Biotechnology; Rockford, IL).
Quantitative RT-PCR and 5' RACE
Total RNA was extracted using RNeasy Mini Kit (Qiagen, Valencia, CA) and converted to cDNA from 2 μg total RNA by SuperScript RT II (Invitrogen, Carlsbad, CA). Quantitative RT-PCR was performed using an ABI StepOnePlus analyzer (Applied Biosystems, Foster City, CA) with SYBR Green master mixture containing primers for ALCAM (NM_001627) or GAPDH (Additional file
1: Table S1). The start of RNA synthesis was identified using the rapid amplification of cDNA ends (RACE) approach (5'RACE, Invitrogen). Briefly, ALCAM cDNA was synthesized from 3 μg RNA using SuperScript RT II and ALCAM-specific primer AL1 (Additional file
1: Table S1), and purified using SNAP column (Invitrogen), tailed with TdT and amplified by PCR using abridged anchor primer (AAP) and ALCAM-specific primer AL2. PCR product was amplified using a nested ALCAM-specific primer AL3 and universal amplification primer (UAP), cloned into a T-vector and sequenced.
DNA constructs and stable cell lines
Human genomic DNA was amplified by PCR using a common reverse primer and various forward primers truncated at -2600, -1800, -1400, -1200, -1000, -800, -650, -400 and -200 (Additional file
1: Table S1). PCR products were cloned into a promoter-less luciferase vector (pGL3, Promega, Madison, WI) via
Mlu I and
Bgl II, and verified by DNA sequencing. We have previously described construction of a fusion DNA vector expressing ALCAM and enhanced green fluorescent protein (GFP) [
47]. ALCAM-GFP and control GFP vector (Applied Vironomics, Fremont, CA) were transfected into log-phase growing MDA-MB-435 cells using lipofectamine 2000 (Invitrogen). Stable lines of MDA-MB-435-ALCAM-GFP and MDA-MB-435-GFP were selected using G418 (500 μg/ml).
Reporter assays
Cells (8 × 104) were seeded in 24-well tissue culture plates and co-transfected with ALCAM promoter luciferase plasmids (800 ng) and pcDNA3.1/His/LacZ (100 ng) (Invitrogen) plasmid DNA using lipofectamine 2000. Twenty-four hours after transfection cell lysates were prepared and the activities of luciferase (Firefly-Luciferase Reporter Assay System, Promega) and β-galactosidase (Galacto-Star system, Applied Biosystems) determined using the Veritas Luminometer (Turner Biosystems, Sunnyvale, CA). Luciferase activity was normalized to the activity of β-galactosidase, and the relative luciferase activity for test constructs calculated by assigning the normalized luciferase activity of the promoter-less pGL3 construct as 1.0. Minimum of three independent experiments was performed for each reporter each in triplicate.
Electrophoretic mobility shift assay
In vitro protein-DNA interaction was examined using the LightShift Chemiluminescent electrophoretic mobility shift assay (EMSA) kit (Pierce). ALCAM-specific EMSA DNA probes were synthesized, gel purified and biotin labeled (Additional file
1: Table S1). Nuclear extract (4 μg) was combined with biotin-labeled DNA probes in binding buffer containing 50 μg/ml poly(dI-dC). Fifty-fold molar excess of unlabelled DNA probe was added to the binding reaction in competition experiments. Anti-p65 NF-κB antibodies (2 μg) (Santa Cruz) were added to the reaction mixture. Products of the binding reaction were resolved in 6% DNA retardation gel, transferred to a nylon membrane and biotin-labeled complexes detected by chemiluminescence (Fujifilm LAS-1000 imaging system; FujiFilm, Valhalla, NY).
Chromatin immunoprecipitation assay
Protein-DNA cross-linking was performed by fixing 40 million cells with 1% formaldehyde. Nuclei was sonicated on ice in shearing buffer (ChIP-IT; Active Motif, Carlsbad, CA) to obtain chromatin fragments of 100-1000 base pairs, which were pre-cleared with protein G beads (Salmon sperm DNA/Protein G agarose). Pre-cleared chromatin was incubated with anti-p65 NF-κB antibody (Santa Cruz) or non-immune IgG. Immune complexes were precipitated with protein G beads, and the eluate reversed cross-linked in 190 mM NaCl. DNA was purified and amplified by PCR with specific ALCAM primers (Additional file
1: Table S1).
DNA methylation
Genomic DNA was extracted using QIAamp DNA Mini Kit (Qiagen) and 1 μg of this sample treated with sodium bisulfite (EZ Methylation Gold kit, Zymo Research, Orange, CA). Bisulfite-modified DNA (1 μl) was amplified in two separate PCR reactions using primers flanking the interval -256 to -118 of the ALCAM promoter and specific for methylated and unmethylated genomic DNA (Additional file
1: Table S1). For bisulfite sequencing biotinylated reverse primers were used in two separate reactions to amplify two distinct and overlapping DNA fragments in the interval -409 to -62. Biotin-labeled single-stranded PCR products were isolated and pyrosequenced (PSQ™96HS System EpigenDx Biotage, Kungsgatan, Sweden). Methylation status of each CpG site was analyzed individually as T/C SNP using QCpG software (Biotage, Kungsgatan, Sweden).
Isolated perfused lungs
The isolated perfused ventilated rat lung was prepared as we have described previously [
52,
53]. Adult male Sprague-Dawley rats were anesthetized with pentobarbital sodium (60 mg/kg ip), and a catheter was inserted into the trachea, and the lungs were mechanically ventilated. A median sternotomy was performed, and then heparin (60 units) was administered via the left ventricle and allowed to circulate for 3 min. Catheters were placed and secured around the pulmonary artery and the left atrium. Rat lungs were perfused with Earle's balanced salt solution and 4% bovine serum albumin containing tumor cells at a density of 40,000 cells per ml. In some experiments, tumor cells (2 × 10
6) were incubated for 1 hour with anti-ALCAM (1:20 Novocastra or 1:20 GW, GenWay Bioetech) or non-immune IgG(control) prior to perfusion. Tumor cells were perfused for 90 minutes followed by perfusion with Earle's balanced salt solution without tumor cells for 5 minutes.
Immunostaining
Breast cancer cells were fixed with methanol and blocked with 3% normal goat serum, followed by staining with 1/100 dilution of anti-ALCAM (Novocastra Laboratories Ltd), and Alexflour 594 goat anti-rabbit IgG (Molecular Probes, Eugene, Oregon). Nucleus was stained with DAPI (Molecular Probe). Cells were examined by epifluorescence (Nikon TE2000, Nikon Instruments Inc., Melville, NY). Lung tissue was fixed in 10% formalin or 4% paraformaldehyde, processed, embedded in paraffin, and sectioned (4-5 microns). Sections were stained with hematoxylin & eosin (H&E), anti-GFP (1:250; Molecular Probes, Inc., Eugene, OR), or anti-ALCAM (1:40; Novocastra Laboratories Ltd). Tumor cells in twenty 40× fields were counted on each section/slide. A total of forty lung fields (40×) per rat lung were analyzed by light microscopy to assess for tumor cells. The number of tumor cells and presence of single cells or cell aggregates were recorded. Sections were photographed with a Nikon E600 light microscope with digital imaging (Nikon Instruments Inc., Melville, NY).
Statistics
The data are reported as the means ± SE for at least three independent experiments. Data was graphed and analyzed using Prism software (GraphPad Software). Statistical analysis of the raw data was performed by two tailed t tests. Differences were considered significant if p value were < 0.05 (*), <0.01 (**) and <0.001 (***).
Acknowledgements
We are grateful to Dr. Warner C. Greene for the NF-κB expression vector, Dr. R Fillmore for guidance with the ChIP assay, and Ms. R Cochran for technical assistance.
This work was supported by NIH grants R01HL077769 and P20MD002314-030001 awarded to S.F Ofori-Acquah and AHA grant 0655377B (J.A. King).
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
JAK participated in the design of the studies and performed experiments. FT performed experiments and prepared the manuscript. FM, ZC, SC, HC, DA and LAS performed experiments. SFOA designed the study, analyzed data and wrote the manuscript. All authors read and approved the final manuscript.