Introduction
Hematopoietic growth factor inducible neurokinin-1 type (HGFIN; also known as nmb) is a single transmembrane protein located in human chromosome 7 [
1,
2]. HGFIN shares sequence similarities with the G-protein, 7-transmembrane coupled neurokinin-1 (NK1) receptor [
1]. This similarity results in HGFIN being able to interact with the high affinity ligand for NK1, substance P [
1]. The murine homolog of HGFIN, osteoactivin, has been reported to upregulate matrix metalloproteinase-3 and -9 in atrophied skeletal muscles from denervation [
3]. Osteoacvtivin is involved in osteoblast development and function [
2,
4,
5]. The fact that osteoactivin is expressed in differentiated osteoblasts is consistent with human HGFIN being linked to differentiated immune cells [
1,
6]. HGFIN has been reported to act as a negative regulator of inflammation [
7,
8]. A recent description of a protein with 99% homology to HGFIN, DC-HIL show that this gene is promotes adhesion in an RGD-dependent manner [
9].
Osteoactivin and HGFIN are widely expressed in normal and malignant cells [
10]. Osteoactivin is expressed in breast cancer cells and has been associated with bone invasion, an aggressive form of the disease [
11].
HGFIN expression is partly regulated by p53 through multiple sites within the 5' flanking region [
12]. In contrast to its expression in cancer cells, in non-transformed cells
HGFIN expression appears to be critical in cell cycle quiescence [
1,
12]. The link between p53 and the regulation of
HGFIN expression leaves the question of the role of
HGFIN in tumor development open. Presence of
HGFIN confers low metastatic potential in melanoma cells [
2].
This study further investigates the regulation of HGFIN, and also determines its involvement in breast cancer. Here, we report on a repressive effect of the non-coding exon 1. We also report on the cause–effect relationship between defect in HGFIN and transformation of non-tumorigenic breast cells.
Materials and methods
Reagents and antibodies
RPMI-1640, DMEM, α-MEM, alkaline phosphatase-conjugated goat anti-rabbit IgG, neurokinin-A, and anti-Flag were purchased from Sigma (St Louis, MO, USA). Fetal calf serum (FCS) and horse serum were purchased from Hyclone Laboratories (Logan, UT, USA).
Cell lines
K562, MCF12A (non-tumorigenic), MCF10 (non-tumorigenic), DU4475 (carcinoma), HCC70 and T47D (low invasive) were purchased from American Type Culture Collection (ATCC; Manassas, VA, USA) and cultured according to their instructions. The highly aggressive cancer cell line MDA-MB-231 was obtained from Dr Ian Whitehead, New Jersey Medical School (Newark, NJ, USA), and was originally described by ATCC. CCL64 has been described previously [
13].
Vectors and reporter gene assay
pGL3-basic and the luciferase detection kit were purchased from Promega (Madison, WI, USA). The β-galactosidase (β-gal) detection kit and pHyg were purchased from Clontech (Palo Alto, CA, USA). pFLAG-CMV2-HGFIN and pPMSKH1 (siRNA) were as previously described [
1,
14]. pCR2.1 was purchased from Invitrogen (Carlsbad, CA, USA). The p53 expression vectors and mutants were kindly provided by Dr Yuzuru Shiio (Institute for Systems Biology, Seattle, WA, USA): pME18S-SN3 wild-type human p53, pME18S-SCX3 contained 143 Val→Ala mutant human p53, and pPME18S [
15]. The vectors are under the SRα promoter. The expression vectors encode both the N- and C-termini of p53 [
15].
Cloning HGFIN-RM/2.0E
The 5' flanking region of
HGFIN, HGFIN-RM/2.0 was as previously reported [
12]. The HUGO Gene Nomenclature Committee has suggested the official symbol of HGFIN as GPNMB: human transmembrane glycoprotein nonmetastatic melanoma protein B. A two-step cloning procedure has been used to add exon 1, with the translational start site omitted, downstream of HGFIN-RM/2.0. The clone has been designated HGFIN-RM/2.0E. The first step used PCR with pooled human gDNA as template and Hot Start Ex
Taq Polymerase (Invitrogen) with the following primers: 5'-ggtgcagggaaggaaaaaagac-3' (sense) and 5'-tagagacattccatgctgaa-3' (antisense). The fragment was inserted into pCR2.1 and was designated HGFIN-RM/2.1. In the next step, we cloned HGFIN-RM/2.0E with primers that include exon 1 with omission of the translation start site: 5'-ctcgaggtgcagggaaggaaa-3' (sense with
XhoI linker) and 5'-aagctttccatgctgaattcc-3' (antisense with
HindIII linker). The fragment was first ligated into pCR 2.1 for sequencing at the Molecular Core Facility, New Jersey Medical School (Newark, NJ, USA). After the DNA sequencing verification, the insert was subcloned into pGL3-basic reporter vector within
XhoI/
HindIII sites.
Transfection and reporter gene assay
Reporter gene assays were performed as described [
16]. Briefly, non-tumorigenic cells at 60–80% confluence were co-transfected with pGL3-HGFIN-2.0 or -HGFIN-2.0E and pβ-gal-control (0.2 μg each). Transfections with Effectene (Qiagen, Valencia, CA) results in 60–80% efficiency as determined by labeling for β-gal [
16]. Controls were transfected with pGL3-basic pβ-gal. After 48h, cell extracts were quantitated for luciferase and β-gal using kits from Promega and Clontech, respectively. The ratios of luciferase/β-gal in cells transfected with vector alone were normalized to 1. Luciferase activities were presented per μg of total protein and the levels normalized with cells transfected with vector alone. Total protein in cell extracts was quantitated using a protein assay kit from BioRad (Hercules, CA, USA).
Stable HGFIN knock-out and expression
The method to construct HGFIN-specific siRNA into pPMSKH1 has been described previously [
14]. HGFIN siRNA was based on NCBI accession number AF322909 spanning +343/+361: 5'-catttgcggtgaacctgat-3'. Blast analyses using NCBI database determined no significant homology to any human gene. The 19 nucleotide sequence (sense) was placed in tandem with the loop structure followed by the antisense sequence of the upstream 19 nucleotide sequence compliment, resulting in 64 nucleotides. Control siRNA contained single nucleotide mutations within the gene-specific insert. Double-stranded DNA was ligated into pPMSKH1 at a molar ratio of 1:50 (vector to insert). Digestion with
EcoRI and
HindIII confirmed inserts of ~280 bp. Negative clones without inserts showed bands at ~220 bp. The insert was further verified by DNA sequencing at the Molecular Resource Facility, New Jersey Medical School, Newark, NJ. Stimulated (GM-CSF) bone marrow fibroblasts have been shown to induce HGFIN and were therefore used to verify the efficiency of siRNA in HGFIN knock-out [
1].
HGFIN knock-out was performed for MCF12A by co-transfecting with pPMSKH1-HGFIN (wild-type or mutant) and pHyg. Stable transfectants were selected with hygromycin at 5 μg/mL. HGFIN expression was studied by co-transfecting T47D with pFLAG-HGFIN and pHyg. Stable transfectants were selected with 150 μg/mL hygromycin. Selected cells were positive for Flag by western blots with combination of whole cell and membrane extracts (data not shown). The combinations of extracts were necessary as HGFIN is a membrane molecule [
1].
Western blots
Cell membrane extracts were obtained as previously described [
17]. Briefly, cells were incubated with 400 μL of 1× lysis buffer (Promega) for 15 min at room temperature. Cell lysates were pelleted by centrifugation at 10,000
g for 15 min at 4°C and the membrane fractions were resuspended in 300 μL of PBS and then vortexed. Whole cells extracts were prepared by repeated cycles of freeze–thaw. The membrane and whole cell extracts were combined and then analyzed for total protein using the BioRad protein assay.
Extracts (10 μg total protein) were electrophoresed on 12% SDS-PAGE. Proteins were transferred to PVDF membranes (Perkin Elmer, Wellesley, MA, USA), and then developed with anti-Flag by overnight incubation at 4°C. After this, membranes were washed and incubated with horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG (1/2,000) for 1 h at 4°C. HRP was developed with chemiluminescence western blot detection reagents (Perkin Elmer). The molecular weights were determined by comparing to Kaleidoscope prestained standards (BioRad).
Cell migration assay
Cell migration was studied in a transwell system with 8.0 μm inserts using 24-well plates (Falcon, Lincoln Park, NJ, USA). Cells (2 × 105) were re-suspended in sera-free media and then added into the inner chamber to a volume of 500 μL. Plates were incubated at 37°C with 5% CO2 for 1.5 h. After this, cells within the inserts were removed with cotton swabs. The filter along with the cells that migrated were fixed and stained with methylene blue. The total numbers of migrated cells were counted with an inverted light microscope (Olympus, Long Island, NY, USA).
Selection of primary breast cancer cells
Breast tissues were obtained at the initial diagnosis of patients with Stages IIIA or IIIB. At the time of surgery, patients were not subjected to chemotherapy or radiation. The use of breast tissues followed the guidelines of the Institutional Review Board, Newark Campus. Patient 7 was obtained from Cooperative Human Tissue Network, University of Pennsylvania Medical Center (Philadelphia, PA, USA). Variations in the hormone status of patients are summarized in Table
1. Malignant cells within the surgical breast tissues were selected as described previously [
18]. The immunohistochemistry analyses were performed with archived samples from the Pathology Department, University Hospital, New Jersey Medical School.
Table 1
Breast cancer study subjects and hormone status.
1–3 | 65–73 | - | - | - | <0.2–1.2 |
19–28 | 54–60 | | | | |
29–35 | 36–60 | | | | |
4–7 | 55–60 | + | + | 1+ | 1–1.5 |
8–10 | 60–64 | | | | |
11–18 | 70–82 | + | + | - | 0.5–1 |
36–50 | 56–65 | | | | |
Benign tissues | 35–60 | Not performed | Not performed | Not performed | 10 |
In situhybridization for HGFIN mRNA
Slides with surgical breast samples from benign and malignant subjects were provided blinded by the co-author of this manuscript (MH). Thus, the status of the patients (benign vs malignant) was not revealed until after the results were obtained and analyzed. All patients were diagnosed with infiltrating ductal carcinomas and none with lobular.
In situ hybridization was performed with a cocktail of three antisense biotinylated oligonucleotides, 18 nucleotides each, specific for the HGFIN mRNA, as previously described [
1,
19]. Briefly, the slides were de-waxed and then incubated with 30 μg/mL proteinase K for 1 h at 37°C. Negative control slides were incubated with 100 μg/ml RNase for 30 min at 37°C. After this, slides were prehybridized with 200 ng/mL oligonucleotide cocktail, each with biotin conjugated at the 5' ends. The oligonucleotides were selected from the two ends, and middle regions of HGFIN cDNA, accession number AF322909 [
1]: 5'-ccacttgatgccgccaaa-3' (+111/+128); 5'-atggcaccggccaaagcc-3' (+496/+513); 5'-gcctgtggtatgatgtgc-3' (+2235/+2252). Sections were then incubated for 1 h at room temperature with 1.25 μg/ml avidin-AP (Boehringer Mannheim Biochemicals). Control slides were incubated with a cocktail of sense oligomers. Slides were counterstained with Harris Modified Hematoxylin (Fisher Scientific, Springfield, NJ, USA) and then examined microscopically with Olympus AX-70 microscope and a Magnafire digital camera (Olympus) as described previously [
20]. Photomicrographs were imported into an image analysis program (analySIS, Soft Imaging System, Munster, Germany) and analyzed to count the positive labelings. Labeling intensities <0.05 were considered negative. The densities of labeling from non-tumorigenic cells were normalized to 10.
Northern analysis
Northern analysis for steady state HGFIN mRNA was performed as described previously [
21]. In brief, total RNA (10 μg) were analyzed with HGFIN cDNA probes, labeled with [α-
32P]-dATP, 3,000 Ci/mM, (Dupont/NEN, Boston, MA, USA). Membranes were stripped and then re-probed with cDNA for 18S rRNA. Probes were randomly labeled with the Prime-IT II random primer kit (Stratagene, La Jolla, CA, USA). Hybridized membranes were developed in a phosphoimager cassette (Molecular Dynamics, Sunnyvale, CA, USA) and then scanned after 16 h on a PhosphoImager (Molecular Dynamics). cDNA for 18S rRNA was purchased from ATCC.
Semi-quantitative RT-PCR
Total RNA was extracted from cells and 2 μg was reverse transcribed. cDNA (200 ng) was subjected to PCR for HGFIN using primers designed from accession number AF322909, spanning +570/+681: 5'-aaccttttcctcaccaccc-3' (forward) and 5'-ttcacagaaactctcactgaac-3' (reverse). PCR reactions were normalized by amplifying the same sample of cDNA with primers specific for glyceraldehyde-3-phosphate dehydrogenase (GAPDH). The primers for GAPDH spanned +212/+809 (NM_002046): 5'-ccacccatggcaaattccatggca-3' (forward); 5'-tctagacggcaggtcaggtccacc-3' (reverse). PCR was performed for 35 cycles for HGFIN and 30 cycles for GAPDH at 94°C for 30 s, 55°C for 30 s and 72°C for 30 s with a final extension at 72°C for 10 min. PCR reactions (10 μL) were separated by electrophoresis on 1.0% agarose containing ethidium bromide. Band sizes were compared with 1 kb DNA ladder (Invitrogen).
Growth curve
Cells were plated at 100 cells/T25 tissue culture flasks. At weekly intervals, cells were trypsinized and then counted.
Methylcellulose cultures
Clonogenic assays were performed as described previously [
14]. Briefly, cells were resuspended in 1.2% methylcellulose containing the respective culture media. Assays were performed with cells seeded at 10
2/ml in 35 mm suspension dishes. Colonies with >20 cells were counted after 5-day incubation at 37°C.
Data analyses
Statistical evaluations of the data were performed with analysis of variance and Tukey-Kramer multiple comparisons test. A result of p < 0.05 was considered significant.
Discussion
This study expands on previous reports that link
HGFIN and its rodent homolog,
osteoactivin, to malignancy. We screened primary breast tissues and found
HGFIN expression in non-tumorigenic cells, but low to undetectable expression in malignant cells (Figure
1; Table
1). The fact that
HGFIN is regulated by multiple p53 binding sites, combined with other studies linking the human gene to low metastatic potential in melanoma cells suggest that
HGFIN might function as a tumor suppressor [
2,
12]. Similar to most genes with tumor suppressor activity,
HGFIN and
osteoactivin have also been linked to properties consistent with malignancies [
4,
11,
25]. Our laboratory has just begun to focus on
HGFIN as a tumor-associated gene, indicating an early investigational stage in studies linked to this gene. To fully understand a role for HGFIN in breast tissues, research studies beyond those presented are required. These include quantitative studies to determine if HGFIN levels are linked to the status of the cancer cells. This could be accomplished in follow-up studies with patients' samples in longitudinal studies, or with isogenic breast cancer cells. It is interesting that
HGFIN is located on chromosome 7, which is surrounded by microsatellite regions. Thus, it would be of interest to examine malignant cells for loss of the
HGFIN gene or loss of heterozygosis.
Exon 1 appears to be critical in the control of
HGFIN expression mainly due to being partly inhibitory in the enhancing function of its activities (Figure
5a,b). It is interesting that the most aggressive cell line, MDA-MB-231 cannot activate
HGFIN reporter gene activity unless p53 was expressed, suggesting that p53 might be the limiting dysfunction in some cancers with respect to HGFIN expression (Figure
5b,c). In addition to the multiple p53 sites reported for HGFIN-RM/2.0, a consensus sequence has been found in exon 1 [
12]. Ectopic expression of p53 led to the activation of HGFIN-RM/2.0E (Figure
5c). However, it is unclear if this increase involves exon 1, as a significant increase was observed for HGFIN-RM/2.0, which has exon 1 omitted (Figure
5c). Exon 1 could be important in unraveling the role of HGFIN in malignancies, not only in breast but also in other cancers. The molecular analysis of HGFIN is the subject of intense research investigation in our laboratory.
We have observed an inverse relationship between
HGFIN reporter gene activity and the aggressiveness of breast cancer cells (Figure
5b) [
26]. The reporter gene activities are consistent with decreased HGFIN mRNA in cancer cells, as compared to non-tumorigenic cells (Figure
1). Although we have shown knock-out of
HGFIN causes an increase in cell growth, contact independent growth and migration (Figures
2,
3,
4), its role needs to be examined with robust genetic approaches. Indeed, computer analyses have shown evidence of HGFIN within a region of microsatellites, which is linked to instability (data not shown). This observation is currently under investigation, with pairs of autologous samples to show whether loss of
HGFIN might be an early event in breast cancer transformation.
In summary, the
HGFIN (or
nmb) gene and its murine homolog,
osteoactivin are unexplored in cancer biology and in particular in the capacity of oncogenes. This study has begun the further examination of this gene at the genetic level. Networks comprising HGFIN with cell cycle regulators, established oncogenes and tumor suppressors need to be elucidated. The location of this gene and its control via multiple p53 sites is intriguing, and might have a critical role in tumor biology. The functional behavior of
HGFIN is reminiscent of the dual role of
p53 as tumor suppressor and as an oncogene [
27,
28]. Finally, the hormone status of patients appears to be irrelevant to the functions of HGFIN, suggesting a global function of HGFIN (Table
1).
Conclusion
HGFIN exhibits properties that are consistent with tumor suppressor gene functions. In its absence, non-tumorigenic cells show evidence of transformation and loss of contact dependency as well as increased migration. These findings have been verified with primary breast tissues in which benign tissues show expression of HGFIN, whereas malignant tissues shown no evidence of HGFIN. The relationship between mutated p53 and HGFIN expression in malignancy of breast cancer and bone invasion will begin to unravel a new pathway used by p53 in breast cancer biology. Also, as this study was performed by overexpression of one variant of HGFIN, it is unclear how the extra 12 amino acid insert between exons 1 and 2 in the extracellular domain of the other human variant will affect the biology reported in this study [29].
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
RLM performed all the assays and prepare a draft of the manuscript; PSP and MB isolated and cultured the primary human breast cancer cells and assisted in preparing the manuscript; MH read the immunohistochemical slides (blinded) and assisted in preparing the manuscript; PR formulated the concept, designed the experiments, and prepared the final manuscript.