Abstract
The nuclear matrix is a dynamic RNA-protein complex that organizes chromatin and regulates nuclear DNA metabolism. Nuclear matrix proteins informative in the diagnosis of cancer have been identified. Here, the nuclear matrix breast cancer proteins (NMBCs) cross-linked to nuclear DNA in situ with cisplatin in human breast cancer cell lines were analyzed by two-dimensional gel electrophoresis. We identified NMBCs that were differentially associated with nuclear DNA of hormone-dependent and -independent breast cancer cell lines. Three DNA cross-linked NMBCs were found to be exclusive to estrogen receptor-positive, hormone-dependent breast cancer cells, whereas two NMBCs were observed only in estrogen receptor-negative, hormone-independent breast cancer cells. Changes in these NMBCs were observed when hormone-dependent breast cancer cells became hormone independent. Furthermore, we show that the intermediate filament protein vimentin is associated with the nuclear DNA of MDA-MB-231 breast cancer cells, an estrogen receptor-negative, hormone-independent breast cancer cell line with high metastatic potential. These nuclear matrix DNA-binding proteins may play important roles in breast tumorigenesis.
INTRODUCTION
Nuclear DNA is organized into loop domains through interactions between MARs3 (DNA sequences at the base of the loop) and nuclear matrix proteins, such as lamins, SatB1, SAF-B/HET, and ARBP (1 , 2) . Aberrant expression of nuclear matrix, MAR-binding proteins could result in the reorganization of chromatin in cancer cells (3) .
The nuclear matrix is a component of the tissue matrix system, which also consists of the cytoskeleton and extracellular matrix. The interplay between these dynamic networks affects nuclear structure, gene expression, and cell phenotype (4, 5, 6, 7) . Using the cross-linker cis-diamminedichloroplatinum (II) (cisplatin), we found that cytokeratins, which are intermediate filament proteins, are associated with nuclear DNA in breast cancer cells. In ER-positive, hormone-dependent breast cancer cells, the interaction between cytokeratins and DNA was modulated by estrogens; however, in ER+, hormone-independent breast cancer cells, this interaction was no longer regulated by estrogens (8) . Thus, both nuclear matrix proteins and intermediate filament proteins appear to have roles in the organization of nuclear DNA.
In addition to serving a role in the organization of chromatin, the nuclear matrix provides a platform for the assembly of protein machines (e.g., replication and transcription factories) involved in the processing of the genetic information (9 , 10) . Proteins and enzymes involved in transcription, chromatin modification, replication, and RNA splicing are targeted to specific nuclear sites, and the nuclear matrix targeting sequences of transcription factors have been identified (11, 12, 13, 14, 15) .
The organization and composition of the nuclear matrix is dynamic, changing with nuclear activity (16, 17, 18) . Some nuclear matrix proteins are found in most cell types (e.g., heterogeneous ribonuclear proteins), whereas others are cell-, differentiation-, or tumor-specific (19 , 20) . The latter group of nuclear matrix proteins can distinguish a malignant cell from a benign or normal cell. Thus, the identification of these nuclear matrix biomarkers has potential in the diagnosis and prognosis of cancer (21) . In an analysis of nuclear matrix proteins from human breast cancer cell lines, we identified nuclear matrix proteins that were found in either well-differentiated or poorly differentiated human breast cancer cells (22) .
In this study, we used in situ cross-linking by cisplatin to identify nuclear proteins that were differentially associated with nuclear DNA in hormone-dependent and -independent human breast cancer cells. Analyses of two-dimensional gel patterns of proteins cross-linked to nuclear DNA revealed that many of the proteins cross-linked to DNA in situ with cisplatin were common to hormone-dependent and hormone-independent cell lines. However, proteins differentially cross-linked to nuclear DNA in these two groups of cell lines were identified. Parallel analysis of two-dimensional gel patterns of nuclear matrix proteins isolated from the cell lines showed that the most proteins differentially cross-linked to DNA were nuclear matrix proteins.
MATERIALS AND METHODS
Cell Culture.
ER+, hormone-dependent (MCF-7, T-47D, T-47D5, and ZR-75), ER+, hormone-independent (T5-PRF), and ER−, hormone-independent (MDA-MB-231, MDA-MB-468, and BT-20) human breast cancer cell lines were used in this study. MCF-10A1, an ER−, hormone-independent, spontaneously immortalized breast epithelial cell line developed from a reduction mammoplasty breast tissue sample was used as a control. Culturing conditions for the cell lines were as described previously (22) .
Isolation and Analysis of Nuclear Matrix Proteins and Proteins Cross-linked to DNA in Situ.
Nuclear matrices were isolated from the breast cancer cell lines as described previously (22) . Intermediate filament proteins were removed from the nuclear matrix protein preparation (23) . Analysis of the nuclear matrix proteins of each cell line was done at least three times. DNA-protein cross-linking was performed as described previously (8) . Briefly, MCF-7 cells at a density of 1 × 106 cells/ml were resuspended in Hanks’ buffer containing sodium acetate instead of NaCl at the same concentration. The cells were incubated with 1 mm cisplatin at 37°C for 2 h with gentle shaking. Following this incubation, cells were treated with lysis buffer (5 m urea, 2 m guanidine-HCl, 2 m NaCl, and 0.2 m potassium phosphate, pH 7.5). Hydroxylapatite (4 g/20 A260 units of lysate, Bio-Rad, Richmond, CA) was then added. The hydroxylapatite resin was washed with lysis buffer to remove RNA and proteins not cross-linked to DNA. To reverse the cross-linking, the hydroxylapatite was incubated in lysis buffer containing 1 m thiourea instead of 5 m urea. By doing so, the proteins were released from hydroxylapatite, while the DNA remained bound. Three separate cross-linking experiments were performed for each cell line studied on cells grown from three consecutive passages. Only proteins that consistently appeared in all three cross-linked profiles of a cell line were considered representative DNA-binding proteins.
Two-dimensional PAGE was performed as described previously (24) . Gels were stained with silver using the Amersham Pharmacia Biotech Plus One Silver Staining Kit, Protein and then dried between sheets of gel drying film (Promega Corp.) at room temperature. Stained gels were scanned using a PDI 325OE densitometer (PDI, Huntington Station, NY), and the data were analyzed with Image Master software (Amersham Pharmacia Biotech).
Immunoblot Analyses.
Proteins electrophoresed on SDS-polyacrylamide gels were transferred to nitrocellulose (Bio-Rad) as described previously (25) . The nitrocellulose membranes were then immunochemically stained with anti-vimentin antibodies, followed by goat anti-rabbit antibodies (Bio-Rad) conjugated to horseradish peroxidase. The immunochemical staining was detected using enhanced chemiluminescence detection system (Amersham Pharmacia Biotech).
RESULTS
ER+, hormone-dependent (T-47D, MCF-7, and ZR-75), ER+, hormone-independent (T5-PRF), and ER−, hormone-independent breast cancer cell lines and an ER−, hormone-independent, spontaneously immortalized, nontumorigenic human breast epithelial cell line (MCF-10A1) were incubated with cisplatin. Proteins cross-linked to nuclear DNA were isolated and analyzed by two-dimensional gel electrophoresis. Representative silver-stained two-dimensional gel patterns of proteins cross-linked to DNA of the various cell lines are shown in Figs. 1⇓ and 2⇓ (see Fig. 3⇓ for a schematic of the two-dimensional gel pattern data). To compare the proteins in the various gel patterns, several exogenous proteins were used to align protein patterns. Carbamylated carbonic anhydrase (30 kDa, pI 4.8–6.7) served as a reference for determining the molecular mass and isoelectric point of the proteins. Two-dimensional SDS-PAGE standards were used to determine the molecular mass of the proteins.
Proteins cross-linked to DNA by cisplatin in situ from ER+ hormone-dependent and -independent breast cancer cell lines. Eighty μg of DNA cross-linked proteins from cells treated with 1 mm cisplatin were electrophoretically resolved on two-dimensional gels. The gels were stained with silver. ca, position of the carbamylated forms of carbonic anhydrase. The position of the molecular mass standards (kDa) is shown to the left of the two-dimensional gel patterns. LA and LC, lamin A and C, respectively. cK8, cK18, and cK19, cytokeratins 8, 18, and 19, respectively. hK, transcription factor hnRNPK. R1 and R2, which are two proteins found cross-linked to nuclear DNA in all cell types tested, serve as internal reference markers.
Proteins cross-linked to DNA by cisplatin in situ from ER−, pseudonormal breast epithelial cell line and hormone-independent breast cancer cell lines. Eighty μg of proteins cross-linked to DNA with 1 mm cisplatin were electrophoretically resolved on two-dimensional gels. The gels were stained with silver. The position of the molecular mass standards (kDa) is shown to the left of the two-dimensional gel patterns. For definitions of ca, LA, LC, hK, cK8, cK18, cK19, R1, and R2 see legend to Fig. 1⇓ . BN, proteins cross-linked to DNA in pseudonormal but not cancer cells. V, vimentin.
A schematic representation of the two-dimensional gel pattern data of proteins cross-linked to nuclear DNA by cisplatin in pseudonormal breast epithelial and breast cancer cell lines. The position of the molecular mass standards (kDa) is shown to the left of the two-dimensional gel patterns. For definitions of ca, LA, LC, hK, cK8, cK18, cK19, R1, R2, and BN, see legends to Figs. 1⇓ 2⇓ .
The gel patterns of proteins isolated from the various cell lines were similar, providing evidence that many of these proteins cross-linked to DNA were in common. A comparison of the isoelectric point and molecular mass co-ordinates of the common DNA-cross-linked proteins with the co-ordinates of proteins identified in previous studies showed that the transcription factor hnRNPK (60 kDa; pI 5.2–5.4) was cross-linked to DNA in all cell lines studied. Two proteins (labeled R1 and R2 in the figures) were consistently observed in the gel patterns. Lamins A and C were also seen in the gel patterns when higher loads of protein were used. In addition, cytokeratins 8 (54 kDa, pI 5.4), 18 (45 kDa, pI 5.3), and 19 (41 kDa, pI 4.9) were identified in the various two-dimensional gel patterns.
Despite the large degree of similarity in the gel patterns of proteins cross-linked to DNA, differences in the abundance of proteins in the two-dimensional gel patterns of proteins isolated from pseudonormal breast epithelial, hormone-dependent, and hormone-independent breast cancer cell lines were observed. In comparing protein levels, hnRNPK was used as an internal loading control. The relative abundance of this protein appeared to be similar in the gel patterns. Table 1⇓ details the proteins displaying differences in silver staining intensity of at least 5-fold. The molecular mass and isoelectric point values of these proteins are summarized in Table 1⇓ . Three DNA-cross-linked proteins were detected only in the pseudonormal MCF-10A1 breast epithelial cell line (BN1, BN2, and BN3), whereas three proteins (NMBC7, NMBC8, and NMBC9) were present at significantly higher levels in the hormone-dependent cell lines MCF-7, T-47D, and T-47D5 when compared to levels observed in other cell line preparations. Hormone-independent, ER− human breast cancer cell lines also displayed two DNA-cross-linked proteins (NMBC10 and NMBC11) that were present at undetectable levels in hormone-dependent or pseudonormal breast epithelial cell lines.
Proteins found in pseudo-normal breast epithelial and breast cancer cell lines
Proteins cross-linked to nuclear DNA of pseudonormal breast epithelial cells but absent in breast cancer cells are labeled as BN. Proteins cross-linked to nuclear DNA of breast cancer cells but not to nuclear DNA of pseudonormal breast epithelial cells are labeled NMBC. The absence or presence of a protein and the level of the protein among the proteins cross-linked to nuclear DNA or associated with the nuclear matrix are indicated by −, +, and ++.
The DNA-cross-linked proteins isolated from the ER+, hormone-independent, T5-PRF breast cancer cell line were compared to those of the parent cell line (T-47D5; ER+, hormone-dependent) from which the hormone-independent cell line was obtained. The two-dimensional gel pattern of DNA-cross-linked proteins isolated from T-47D5 cells was similar to that of the MCF-7 DNA-cross-linked proteins with the exception that lamins A and C were much more prominent in the T-47D5 protein pattern (data not shown). As with the MCF-7 protein pattern, DNA-cross-linked proteins NMBC7, NMBC8, and NMBC9 from T-47D5 cells were at greater levels than those of pseudonormal (MFC-10A1) and ER− cell lines. In contrast, the levels of these three DNA-cross-linked proteins were considerably reduced in the T5-PRF cell line (Table 1)⇓ .
To determine whether the DNA-cross-linked proteins listed in Table 1⇓ were nuclear matrix proteins, we inspected the two-dimensional gel pattern of nuclear matrix proteins from the respective cell line for the presence of the DNA-cross-linked protein (22) . This comparison showed that all but two (BN1 and BN2) of these DNA-cross-linked proteins were nuclear matrix proteins (Table 1)⇓ . However, these DNA-cross-linked proteins appeared to be present at similar levels in the various nuclear matrix protein preparations in which the intermediate filament proteins had been removed (data not shown).
In a previous study, five nuclear matrix proteins, designated NMBC1–NMBC5, were found to be more abundant in well-differentiated cell lines (T-47D, MCF-7, ZR-75, and T5-PRF), whereas NMBC6 was present at higher levels in poorly differentiated breast cancer cell lines (22) . Among the DNA-cross-linked proteins, NMBC1 (57 kDa, pI 5.5) and NMBC2 (62 kDa, pI 5.1) were the most prominent in all of the cell lines studied, including MCF-10A1. NMBC3 (40 kDa, pI 5.4), NMBC4 (41 kDa, pI 5.3), NMBC5 (39 kDa, pI 5.5), and NMBC6 (52 kDa, pI 5.7) were not detected in the two-dimensional profiles of DNA-cross-linked proteins.
Previously, we had found that the intermediate filament proteins cytokeratins 8, 18, and 19 of human breast cancer cell lines were cross-linked to nuclear DNA by cisplatin (8) . Among these cell lines, MDA-MB-231 cells also express the intermediate filament protein vimentin. The MDA-MB-231 preparation had a protein with a molecular mass and pI (57 kDa, pI 5.0) similar to that of vimentin. To provide evidence that vimentin was cross-linked to nuclear DNA in situ, the proteins were analyzed by immunoblotting with anti-vimentin antibodies. For comparison, a preparation of high-salt Triton-insoluble proteins enriched in cytoskeletal proteins was included in the analysis. Fig. 4⇓ shows that vimentin was cross-linked to nuclear DNA in situ in MDA-MB-231 cells with cisplatin. As a control, proteins cross-linked to DNA of MDA-MB-468 cells, which do not express vimentin, were also analyzed by immunoblotting with anti-vimentin antibodies. Immunoreactive protein was not detected in this sample, showing that the antibody was specific for vimentin (data not shown).
Association of vimentin with nuclear DNA of a metastatic ER− human breast cancer cell line. Cisplatin DNA-cross-linked (X-LINK) proteins (45 μg) and cellular intermediate filament (IF) proteins (45 μg) from the MDA-MB-231 cell line were resolved on a one-dimensional SDS polyacrylamide gel, transferred to a nitrocellulose membrane, and immunochemically stained with anti-vimentin antibody. Lane 1, cellular intermediate filaments. Lane 2, cisplatin DNA-cross-linked proteins. V, vimentin. This is a representative result obtained from three different preparations.
DISCUSSION
The nuclear matrix is involved in the organization and function of nuclear DNA (18 , 26 , 27) . Nuclear matrix proteins that are informative in cancer diagnosis have been identified (21 , 28 , 29) . Previously, we identified nuclear matrix proteins, which were informative in breast cancer (22) . Here, a strategy to identify nuclear matrix proteins that are associated with nuclear DNA in situ was applied to compare pseudonormal breast epithelial with breast cancer cell lines. The nuclear matrix proteins bound to nuclear DNA presumably have a function in chromatin organization and/or function. Thus, the nuclear matrix proteins differentially associated with nuclear DNA in cancer versus normal cells may have a role in chromosome instability, as well as in aberrant nuclear DNA arrangement and/or metabolism in cancer (3 , 30) .
The cross-linker used in this study was cisplatin. This agent is particularly useful in the identification of nuclear matrix proteins involved in the organization of nuclear DNA because it preferentially cross-links nuclear matrix proteins to MAR DNA (23) . Unlike formaldehyde, which will cross-link protein to DNA and protein to protein, cisplatin directly cross-links DNA to protein (cross-link distance, 4 Å; Ref. 31 ).
We identified several NMBC proteins that are differentially associated with nuclear DNA of ER+, hormone-dependent versus ER−, hormone-independent breast cancer cells. As the features of breast cancer cells are similar to that of their corresponding tumors, the NMBCs identified in this study (e.g., NMBC10 and NMBC11) may serve to distinguish breast cancer epithelial cell types present in breast tumors (32 , 33) . The identification of nuclear matrix proteins differentially associated with the chromatin of T5-PRF (ER+, hormone-independent cell line) and parent T-47D5 (ER+, hormone-dependent cell line) is of particular importance in breast cancer progression because most ER+ tumors will eventually become hormone independent (34) . It should be noted that although the T5-PRF cell line was hormone independent, the cells did not assume the phenotype of a ER−, hormone-independent cell line with regards to association of NMBC10 and NMBC11 with nuclear DNA.
We have discussed previously that the conventional method of isolating and identifying cancer-specific nuclear matrix proteins may miss informative nuclear matrix proteins (8) . During the preparation of nuclear matrix proteins, intermediate filaments are often removed (24) . Previously, we reported that intermediate filament proteins, cytokeratins, were bound to nuclear DNA of breast cancer cells. Here, we show for the first time that vimentin is in contact with nuclear DNA in the MDA-MB-231 cell line, an ER− human breast cancer cell line with high metastatic potential. An in vitro study by Wang et al. (35) has shown that vimentin binds in a sequence-specific manner to DNA. Vimentin contributes to the metastatic potential of human breast tumors and melanoma (36, 37, 38) . It is conceivable that intermediate filaments composed of vimentin communicate signals from the extracellular matrix to nuclear DNA, resulting in changes in gene expression associated with metastasis (5 , 39) .
In summary, the characterization of nuclear matrix proteins and proteins cross-linked to DNA by cisplatin in situ are two complementary approaches that can be used to identify informative nuclear matrix proteins in cancer diagnosis.
Footnotes
-
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
-
↵1 This research was supported by Medical Research Council of Canada Grant RO-15183 and by the Manitoba Health Research Council. J. R. D. is a Medical Research Council of Canada Senior Scientist.
-
↵2 To whom requests for reprints should be addressed, at Manitoba Institute of Cell Biology, University of Manitoba, 100 Olivia Street, Winnipeg, Manitoba R3E OV9, Canada. Phone: (204) 787-2391; Fax: (204) 787-2190; E-mail: Davie{at}cc.umanitoba.ca
-
↵3 The abbreviations used are: MAR, matrix attachment region; NMBC, nuclear matrix breast cancer protein; ER, estrogen receptor.
- Received July 15, 1999.
- Accepted November 11, 1999.
- ©2000 American Association for Cancer Research.