Osteopontin (OPN) is a secreted phosphoprotein involved in cellular proliferation and associated with tumor progression. Although an intracellular form of OPN has been described, its function remains unknown. In this study, a novel nuclear location for intracellular OPN and a correlation with cell division were demonstrated. OPN distinctly localized to the nucleus in a subset of transiently transfected human embryonic kidney 293 cells. Immunoblotting confirmed the nuclear location of native OPN, and results from immunofluorescence studies suggested an association between nuclear OPN and cell cycle progression. Flow cytometry revealed that nuclear and cellular OPN content rose significantly during the S and G2/M phases, respectively. Treatment of cells with the DNA polymerase inhibitor aphidicolin prevented cell cycling and greatly reduced cellular OPN content. The intracellular location of OPN coincided with polo-like kinase-1 (Plk-1), a member of the polo-like kinase family, which, in part through their regulation of centrosome-related events, are integral to successful cellular mitosis. OPN and Plk-1 were coimmunoprecipitated from nuclear, but not cystoslic, extracts, demonstrating an interaction that is limited to the nucleus, presumably during mitosis. Deletion of the COOH terminus of OPN militated against nuclear localization and Plk-1 interaction. Elevated expression of OPN was also associated with an increase in the number of multinucleate 293 cells, whereas transfection of the COOH-terminal-deleted OPN decreased the percentage of multinucleate cells below basal levels. These findings implicate intranuclear OPN as a participant in the process of cell duplication.
- cell cycle
osteopontin (OPN) is a phosphorylated glycoprotein highly expressed in bone tissue (9), where it regulates mineralization through its actions on bone cell adhesion and osteoclast function (11). OPN is also present in other tissues and is associated with a wide array of cellular processes, including, but not limited to, development, carcinogenesis, immunity, inflammation, and tissue repair (for review see Ref. 4). In nonbone tissues, OPN expression occurs only in response to a stimulus such as inflammation. Secretion of OPN under these conditions has been correlated with increased tissue calcification; however, secreted OPN also operates as a chemoattractant for various cell types, including endothelial cells and monocytes/macrophages. OPN may be the primary factor in the recruitment of macrophages to sites of injury. This function is mediated by the RGD domain of OPN, which is required for integrin binding. OPN has also been reported to influence cell proliferation, differentiation, and activation state, although the specific mechanisms by which it operates have not been completely elucidated.
Tumor growth has been shown to be directly enhanced by OPN, at least partially through its ability to facilitate tumor invasion by invoking matrix-degrading enzymes (18). OPN expression has also been associated with the growth of nonneoplastic cells (10). Inhibition of OPN or antagonism of its αvβ3-integrin receptor abrogates mitogen-induced DNA synthesis in cardiac fibroblasts and NIH 3T3 cells (2, 23). Similarly, glucose- and hypoxia-induced DNA synthesis in mesangial cells is inhibited by OPN neutralization (20). More direct evidence for OPN's mitogenic capacity was provided by its ability to stimulate division of previously quiescent prostatic cells (8). Enhanced expression of OPN is not limited to isolated cells in culture, inasmuch as OPN localizes to actively proliferating cells in animal models of renal, vascular, and periodontal tissue injury and regeneration (16, 17, 27, 29).
In an attempt to identify the contribution of OPN to cell proliferation, we transfected OPN as a fusion construct with green fluorescent protein (GFP) into 293 cells. In addition to being present in the cytosol, the OPN-GFP fusion protein was found to localize to cell nuclei. Examination of native OPN indicated that its nuclear location correlated with chromatin condensation during cell division, exhibiting a subcellular distribution that coincided with polo-like kinase-1 (Plk-1), an established regulator of mitosis (12). Evidence that OPN and Plk-1 coimmunoprecipitate raises the possibility that these proteins may function cooperatively to modulate cell division cycles.
MATERIALS AND METHODS
OPN expression vectors.
The coding region for full-length human OPN was amplified by PCR with primers [ATGAGAATTGCAGTGATTTGCTTTTGC (forward) and TTAATTGACCTCAGAAGATGCACTATC (reverse)] from a human kidney cDNA library (Clontech) and ligated into the mammalian expression vector pcDNA3.1/NT-GFP-TOPO (construct designated GFP-OPN; InVitrogen). In addition, OPN cDNA was cloned out of frame with GFP into pcDNA3.1/NT-GFP-TOPO, providing a vector closely matched in size and capable of producing GFP, but not OPN (construct designated GFP-CON). A COOH-terminal deletion of the last 101 amino acids (bases 601–903) of GFP-OPN (designated GFP-OPN-C) was made with a site-directed mutagenesis kit (Clontech), and its sequence was verified by cycle sequencing. Cycle sequencing was also used to verify that GFP-OPN and GFP-CON constructs would and would not express OPN, respectively, according to proper reading frame orientation in relation to the coding region for GFP.
293 cells were obtained from the American Type Culture Collection and propagated in DMEM containing 10% FBS. Cells were used when 60% confluent. For microscopy, cells were prepared in chamber slides coated with collagen type I. Cell growth was arrested by placement of the cells in DMEM containing 0.25% FBS for 48 h.
Transfection and expression analysis.
Transfection of GFP-OPN, GFP-CON, or GFP-OPN-C was achieved with Effectine reagent according to the manufacturer's suggested protocol (Qiagen). At 48 h after transfection, the cells were washed in PBS, treated with Hoechst 33342 stain to facilitate nuclear identification, and fixed with 4% paraformaldehyde. Sections were examined for fluorescence with an Olympus BH-2 microscope, and images were captured with a Dage-MTI 330 digital camera. Transfection efficiency was determined by dividing the number of fluorescent cells by the total number of cells. All transfections were repeated a minimum of three times.
After treatment, cells were washed with PBS and fixed with 4% paraformaldehyde as previously described (30). Cells were permeabilized with Tris-buffered saline (TBS)-1% Triton X-100 for 10 min, washed with TBS-0.05% Tween 20 (TBS-T), and then incubated with TBS-T-1% BSA for 60 min at room temperature. A 1:100 dilution of monoclonal anti-OPN antibody (catalog no. MPIIIB10, Developmental Studies Hybridoma Bank) or an equal concentration of isotype-specific mouse IgG (Sigma) was then applied for 1 h at room temperature. After the cells were washed with TBS-T, a Cy2- or Cy3-conjugated secondary antibody (Molecular Probes; 1:500 dilution) was applied for 30 min at room temperature. For OPN-Plk-1 double-labeling experiments, a polyclonal anti-Plk-1 antibody (Santa Cruz Biotechnology; 1:100 dilution) was used. Nuclei were visualized by Hoechst 33342 staining and then examined by fluorescence microscopy as described above. All studies were performed in triplicate.
Flow cytometry analysis of the cell cycle.
For intact cell studies, 293 cells growth arrested in the G0/G1 phase were treated, trypsinized, and fixed in 4% paraformaldehyde. The cells were washed and permeabilized for 5 min on ice with 0.25% (vol/vol) Triton X-100 in PBS (pH 7.4) and then incubated with a 1:100 dilution of anti-OPN antibody or an equivalent concentration of isotype-specific mouse IgG in PBS-1% BSA overnight at 4°C with gentle agitation. After they were washed, cells were incubated for 30 min at room temperature with Oregon green-conjugated secondary antibody (1:500 dilution; Molecular Probes), washed with PBS-1% BSA, and incubated with propidium iodide (PI) staining solution (5 μg/ml PI and 200 μg/ml DNase-free RNase A in PBS) in the dark for 20 min.
For nuclear studies, randomly cycling 293 cells were scraped from dish surfaces and suspended in ice-cold nuclear extraction buffer [320 mM sucrose, 5 mM MgCl2, 10 mM HEPES, and 1% Triton X-100 (pH 7.4)] at ∼1 × 106 cells/ml. The suspension was vortexed gently for 10 s and incubated on ice for 10 min. Nuclei were collected by centrifugation at 2,000 g and washed twice with nuclear wash buffer [320 mM sucrose, 5 mM MgCl2, and 10 mM HEPES (pH 7.4)]. Nuclear yield and integrity were confirmed by microscopic examination after trypan blue staining. Freshly isolated nuclei were incubated with a 1:100 dilution of anti-OPN antibody or mouse IgG in nuclear wash buffer overnight at 4°C with gentle agitation. Secondary antibody incubation and PI staining were conducted as described for intact cells, except nuclear wash buffer was used as the diluent.
Flow cytometry was performed using an Epics XL-MCL flow cytometer (Coulter). Fluorescence values obtained from cells or nuclei incubated with nonspecific mouse IgG served as a measure of background fluorescence and were subtracted from fluorescence values of cells incubated with anti-OPN antibody. Intensity of Oregon green fluorescence above background values, which represents the OPN-specific signal, was correlated with cell cycle phase (as determined by PI fluorescence) by analysis with FCAP software (version 1.32, Soft Flow Hungary).
Immunoprecipitation and Western blotting.
Immunoprecipitation and Western blotting were performed as described previously (30, 31). Briefly, aliquots containing equal amounts of nuclear or cytosolic protein (100 μg) were incubated with anti-OPN or anti-Plk-1 antibody (3 μg) overnight at 4°C. Precleared protein G-Sepharose was used to isolate antibody-bound proteins. After the aliquots were washed and eluted from Sepharose, equal aliquots of immunoprecipitates were mixed with bromphenol blue and 2-mercaptoethanol [1% (vol/vol) final concentration] and loaded onto 7.5% polyacrylamide gels after they were heated at 95°C. After electrophoresis, proteins were transferred to polyvinylidene difluoride membranes that were subsequently blocked with PBS-3% (wt/vol) BSA. The membranes were incubated with anti-OPN antibody (anti-Plk-1 immunoprecipitate) or anti-Plk-1 antibody (anti-OPN precipitate) in PBS-1% (wt/vol) BSA (diluted 1:100) at 37°C for 1 h. After the membranes were washed three times and incubated for 1 h at 37°C in blocking solution with horseradish peroxidase-conjugated secondary antibody, the horseradish peroxidase signal was detected using the enhanced chemiluminescence system (Amersham). Nuclear extract purity was confirmed by the absence of cytoskeletal actin (Sigma) and the presence of Oct-1 (Upstate Biotechnology). Both antibodies were used at a dilution of 1:250.
Transfected OPN localizes to the nucleus.
As has been previously described for native OPN (34), expression of an OPN-GFP fusion protein in 293 cells resulted primarily in cytoplasmic fluorescence in the transfectants (Fig. 1A). However, cells that fluoresced in nuclear and cytoplasmic compartments, as well as cells that exhibited predominantly nuclear fluorescence, were also observed (Fig. 1, B and C). The homogeneous distribution between the cytosol and the nucleus after transfection of a GFP that was out of frame with OPN demonstrated that the nuclear OPN localization is not an artifact of the GFP tag (Fig. 1D). Interestingly, noncytosolic OPN-GFP fluorescence appeared greatest within nuclei with condensed chromatin, suggesting a possible relation between intracellular OPN distribution and mitotic bodies.
Native OPN localizes to the nucleus during mitosis.
Since intracellular trafficking of endogenously produced OPN and GFP-OPN fusion protein may be regulated differently, the intracellular distribution of native OPN was studied by immunofluorescence imaging. OPN staining typically revealed an eccentric, nonnuclear, polar distribution in the majority of cells (Fig. 2, A and B). However, consistent with the results of the transfection experiments, nuclei of cells exhibiting chromatin condensation stained strongly for OPN (arrows in Fig. 2, A and B). An examination of these cells at higher magnification suggested a relation between the location of nuclear OPN and specific stages of mitosis: OPN surrounded the condensed chromatin of prometaphase and metaphase figures (white arrows in Fig. 2, C and D), whereas it was present within the cleavage furrow of mitotic cells in anaphase (green arrow in Fig. 2, C and D). To further investigate this relation between mitosis and nuclear OPN, the subcellular localization of OPN was monitored after treatment with aphidicolin, which blocks cell cycle progression into the S phase and, consequently, reduced the number of mitotic figures (data not shown). Nevertheless, OPN fluorescence remained visible in nuclei with condensed chromatin, consistent with mitosis (arrows in Fig. 2, E and F). On the other hand, aphidicolin treatment produced a visible diminution in cytosolic OPN (cf. Fig. 2, B and F). These data suggest that the intracellular distribution of OPN (within nuclear or nonnuclear compartments), as well as OPN synthesis, may correlate with specific stages of the cell cycle.
Cell cycle-dependent expression and localization of OPN.
The relation between OPN expression and distribution with cell cycle progression was examined by flow cytometry. The DNA content of normally cycling 293 cells indicated that the majority of cells were in the G0/G1 phase (61%), with the remainder in the S (14%) or the G2/M (25%) phase (data not shown). Serum deprivation (0.25% FBS) for 48 h greatly diminished S (5%) and G2/M (15%) phase populations and cellular OPN content (Fig. 3, A and B). At 18 h after stimulation with 10% FBS, there was a significant increase in DNA synthesis (25% of cells in the S phase) and a concomitant increase in OPN expression (Fig. 3, C and D). Cell cycle phase distribution and OPN expression profiles had completely normalized by 64 h (Fig. 3, E and F). The relation between DNA replication and OPN expression was confirmed by the fact that addition of 10% FBS to cells arrested with aphidicolin did not induce appreciable expression of OPN after 64 h (Fig. 3, G and H).
The data presented in Fig. 3 suggest that the appearance of intracellular OPN is dependent on cell cycle stage, with expression primarily associated with the S and G2/M phases. To determine whether the subcellular localization of OPN is similarly modulated in relation to the cell cycle, flow cytometry was used to measure total cellular and nuclear-specific fluorescence in the G0/G1, S, and G2/M phases. The fluorescence intensity of cellular OPN was significantly greater in the G2/M phase than in the G0/G1 and S phases (Fig. 4, A and B). Similarly, nuclear OPN was more intense in the G2/M than in the G0/G1 phase (Fig. 4, C and D); however, nuclear OPN was also significantly higher in the S than in the G0/G1 phase. Nevertheless, OPN intensity was greater in the G2/M than in the S phase. The purity of the nuclei used in this experiment was established by Western blotting. OPN and the constitutive transcription factor Oct-1 were present (Fig. 4E), but actin could not be detected (data not shown).
Exportin-1 (CRM1)-dependent nuclear export of OPN.
The possibility that nuclear transit of OPNs occurred via the exportin-1 pathway was examined with the selective inhibitor leptomycin B. Accumulation of OPN in nonmitotic nuclei of proliferating 293 cells was observed by 1 h after addition of leptomycin B relative to control (Fig. 5, A and B). By 2 h, the majority of nuclei stained strongly for OPN (Fig. 5C). Nuclear cyclin B1 content, which is reduced during interphase largely due to exportin-1-mediated exit (13), similarly increased during the observation period (Fig. 5, D–F). These data indicate that nuclear accretion of OPN is likely controlled by modulation of protein export, rather than import.
Interaction of OPN with Plk-1.
The distinct pattern of OPN distribution in nondividing and dividing cells is highly analogous to that seen with the polo-like kinases, which oscillate between polar and mitotic spindle locations (12). Thus, given the similar spatial and temporal distribution of these proteins, we investigated the possibility that Plk-1 and OPN were physically associated. In nondividing cells, Plk-1 and OPN showed an asymmetric distribution with close proximity to the nucleus (Fig. 6, A and B). Colocalization of these proteins was evident when the images were superimposed (Fig. 6C). This spatial overlap was maintained during cell division, where OPN and Plk-1 colocalized to regions consistent with centrosomes and the midzones of mitotic spindles (Fig. 6, D–F). The position of the centrosome was confirmed by staining for γ-tubulin (data not shown).
Although immunostaining indicates that OPN and Plk-1 may associate with the centrosome and mitotic spindle at similar time points during the cell cycle, this technique provides no information with respect to a physical interaction. When nuclear extracts immunoprecipitated with anti-Plk-1 antibody were examined by Western blot analysis, OPN could be detected (Fig. 7). Conversely, Plk-1 could be immunoprecipitated from the nuclear extracts with anti-OPN antibody (Fig. 7). However, coprecipitation did not occur when a cytosolic fraction, instead of the nuclear extract, was used (data not shown). Consequently, the potential physical interaction between these proteins is likely restricted to the nucleus during periods of cell division.
To determine whether the physical interaction of OPN and Plk-1 is a determinant of spatial distribution, we examined the effect of GFP-OPN-C on their association. Cells were transfected with GFP-OPN or GFP-OPN-C, and their subcellular distribution was compared with that of Plk-1. GFP-OPN was prominent in the nuclei of cells exhibiting condensed chromatin (Fig. 8, A and B). Immunostaining for Plk-1 showed a similar distribution (Fig. 8C). In contrast, GFP-OPN-C was dispersed through the cytosol, although it was most abundant in cells with condensed chromatin (Fig. 8, D and E). Plk-1, in contrast, retained its nuclear proximity. These results support the view that entry of OPN into the nucleus may be the outcome of a physical interaction with Plk-1.
Association of OPN expression with multinucleation.
Plk-1 regulates spindle formation and cytokinesis, and interference with its activity results in mitotic aberrations (12). We therefore examined 293 cells to determine whether expression of OPN resulted in visible mitotic defects. A significant number of multinucleated 293 cells were observed under normal culture conditions (Fig. 9A), in accordance with previous reports indicating that multinucleation is a common event in transformed cells (15, 21). No change in multinucleation could be discerned in 293 cells after transfection with empty vector; however, there was a sixfold increase in the cells that were transfected with GFP-OPN (Fig. 9B). Interestingly, there was a 67% decrease in basal multinucleation in cells transfected with GFP-OPN-C (Fig. 9B). These data suggest that translocation of OPN to the nucleus may interfere with proper spindle formation by Plk-1.
In this report, we have demonstrated by complementary techniques, that OPN accumulates in the nuclei of 293 cells via an exportin-1-dependent mechanism. Entry of OPN into the nucleus coincides with the S phase and precedes significant increases in cellular OPN content in the G2/M phase. Furthermore, progression through the S phase was required for intracellular OPN expression, since inhibition of DNA replication with aphidicolin correlated with low levels of OPN. It was also shown that OPN interacts with Plk-1 and that this association may be necessary for retaining OPN in the nucleus. Finally, we show that overexpression of OPN correlates with multinucleation, which suggests that OPN likely affects spindle formation by Plk-1.
Most of OPN's actions at the cellular level have been attributed to the extracellular effects that occur after secretion, where receptor binding results in the activation of specific intracellular signaling pathways. Indeed, OPN receptors include αv (β1, β3 or β5) and (α4, α5, α8 or α9) β1-integrins and variants 6 and 7 of the hyaluronan receptor CD44 (5). However, evidence for the existence of an intracellular form of OPN localized to the perimembranous region has been accumulating, and a specific role for this protein in cell migration has been reported (33). Furthermore, the inability to internalize secreted OPN (34) indicates that intracellular OPN originates by a distinct mechanism that has yet to be identified.
The intracellular form of OPN is primarily associated with the ERM (ezrin, radixin, moesin) proteins that connect the cortical cytoskeleton with the plasma membrane (34). Furthermore, OPN associates with these proteins only in the presence of CD44 (34). These interactions may explain how intracellular OPN mediates cell migration. However, there is also evidence that suggests that intracellular OPN has other cellular functions, particularly in relation to cell proliferation (8, 25, 26, 28).
A Reinhardt analysis of the amino acid composition of human OPN yielded a reliability score of 94.1, a value strongly supporting the likelihood of a nuclear location for this protein (19). Nevertheless, our data suggest that OPN is detectable within the nucleus only during mitosis. Since mitosis is not a frequent event, it may be presumed that evidence for an OPN presence in the nucleus could have been interpreted as an anomaly by other investigators. Our advantage is likely related to the fact that 293 cells are high proliferative and, therefore, have an abundance of mitotic bodies in a typical heterogeneous population.
Since nuclear OPN content peaked when the S phase was highest, whereas whole cell OPN content did not increase until the G2/M phase, it is likely that nuclear translocation of OPN precedes the increase in cellular OPN content. Thus the mechanisms regulating nuclear OPN content may be similar to those of proteins such as cyclin B and β-catenin, which rely on exportin-1-dependent and -independent nuclear export pathways (7). This interpretation is supported by the nuclear accumulation of OPN in the presence of leptomycin B (Fig. 5).
In agreement with the data demonstrating occupancy of nuclei by OPN only during periods of DNA synthesis, aphidicolin-treated cell nuclei did not stain for OPN (data not shown). Additionally, the finding that proliferating aortic vascular smooth muscle cells suppress OPN expression on cell cycle arrest reinforces the premise that OPN expression is intimately related to not just induction but also maintenance of cell cycle progression (24). These findings are in agreement with those of Zohar and colleagues (33), who noted an increase in fetal rat calvarial cell OPN content during cell cycle progression.
Cell cycle-dependent translocation between the cytosol and the nucleus, particularly during mitosis, is a characteristic of only a few proteins. Postulating a connection to the polo-like kinases was therefore to be expected given our data. However, establishing the existence of a physical association between OPN and Plk-1 is not sufficient to define a function for OPN during mitosis. Polo-like kinases regulate several key aspects of mitotic progression (12), and mutations in Plk-1 have been reported to cause multinucleation in response to a spindle defect that prevents cytokinesis (32). Similar to most transformed cells, 293 cells exhibit a high basal rate of multinucleation. However, 293 cells also express an unusually high background level of OPN. Given the association between OPN and Plk-1 that was observed in this study, it is plausible to reason that OPN might influence that activity of Plk-1. Indeed, OPN has been reported to prevent and promote multinucleation in different cell types (1, 22). Also, osteoclasts and macrophages from OPN-deficient mice contain fewer nuclei (3). Thus the fact that both proteins have been shown to influence multinucleation (Fig. 9) supports our speculation that Plk-1 and OPN may operate in tandem.
Our data point to a unique connection between OPN and cell cycle progression and are consistent with reports of correlating intracellular OPN with squamous cell hyperplasia and carcinoma and the transformation of chondrocytes and small round cells (6, 14). Further examination of the function of intracellular OPN during cell division should serve to provide significant insights into the intricacies of cell cycle progression and tumor biology.
This work was supported by grants from the St. Boniface Hospital Research Foundation and Manitoba Health Research Council (to A. Junaid) and the Canadian Institutes for Health Research (to P. Zahradka). M. C. Moon and G. E. J. Harding were supported by fellowships from the Manitoba Health Research Council.
We thank Xueping Xie and Brenda Wright for excellent technical support.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society