The Bin1 gene encodes a c-Myc-interacting adapter protein with tumor suppressor and cell death properties. In this study, we offer evidence that Bin1 participates in a mechanism through which c-Myc activates programmed cell death in transformed primary chick or rat cells. Antisense or dominant inhibitory Bin1 genes did not affect the ability of c-Myc to drive proliferation or transformation, but they did reduce the susceptibility of cells to c-Myc-induced apoptosis. Protein-protein interaction was implicated, suggesting that Bin1 mediates a death or death sensitization signal from c-Myc. Our findings offer direct support for the “dual signal” model of Myc apoptotic function, based on interactions with a binding protein. Loss of Bin1 in human tumors may promote malignant progression in part by helping to stanch the death penalty associated with c-Myc activation.
One of the central features of neoplastic pathophysiology is dysfunctional apoptotic signaling. For example, to support oncogene activation, cells must evolve resistance to intrinsic apoptotic governors that restrict inappropriate cell division. c-Myc is an important human oncogene that has an intrinsic cell death penalty associated with its activation. However, malignant cells in which c-Myc is activated elude this death penalty at some level. How c-Myc drives cell death in malignant cells and how such cells elude this process are questions of significant interest. Recent work indicates that the p53 pathway plays an important role in c-Myc-mediated apoptosis, particularly in mesenchymal cells, but there is also sound evidence that c-Myc can trigger apoptosis through p53-independent mechanisms, particularly in epithelial cells (1).
The Bin1 gene encodes a c-Myc-interacting adapter protein that has been implicated in tumor suppression and cell death processes in malignant human cells (2, 3). Bin1 frequently is down-regulated or inactivated in breast and prostate cancers and in malignant melanoma, and its ectopic reexpression engages a programmed cell death process (2, 3, 4, 5, 6). We explored the role of Bin1 in apoptosis further using two classical primary cell systems for the analysis of c-Myc function, namely chick fibroblasts and BRK4 epithelial cells. Our findings suggest that Bin1 participates in a death mechanism engaged by c-Myc in neoplastically transformed primary cells. Thus, the epigenetic loss or inactivation of Bin1 that frequently occurs in human tumor cells may help those cells elude the death penalty associated with c-Myc activation.
MATERIALS AND METHODS
Plasmid and Retrovirus Construction.
RCAS (BP) (A and B envelope subtypes), and RCOS (BP) (A envelope subtype) are related chicken retroviral vectors that are competent for replication (kindly provided by S. Hughes NCI-FCRDC, Frederick, MD). RCAS (BP) contains the avian leukemia virus LTR, which is ∼10-fold more active than the RAV-0 LTR in RCOS (7). The MBD used, termed ATG99, has been described (2). This domain is sufficient for interaction with c-Myc and dominantly interferes with c-Myc-Bin1 interaction in vivo (2). The MBD was subcloned into RCAS (BP)-B to generate RCAS-MBD. The dominant inhibitory mutant Bin1 DN was generated by standard PCR methodology as part of a structure-function analysis of Bin1 (8). Bin1 DN lacks amino acids 143–148 of human Bin1 (2). For expression in chick cells, MBD and Bin1 DN cDNAs were subcloned into RCAS (BP)-B via the adapter cloning plasmid pCla12-Nco (9). A human c-Myc cDNA subcloned into pCla12-Nco was subcloned into RCAS (BP)-A or RCOS (BP)-A. An RCAS-Bcl-2 vector has been described (10). For REF transformation experiments, Bin1 DN was subcloned into the CMV enhancer/promoter-driven vector pcDNA3 (Invitrogen). The H-ras vector pT22, human c-Myc vector LTR Hm, and Bin1 vector CMV-Bin1 each have been described (2). For expression in BRK epithelial cells, MBD and Bin1 DN cDNAs were subcloned into MSCVneoEB, a Moloney murine leukemia virus retroviral vector that includes a neomycin resistance gene cassette (11).
Tissue Culture and Cell Line Generation.
Cells were cultured in DMEM containing 10% FCS (Life Technologies) and penicillin/streptomycin unless otherwise indicated. Primary CEFs were cultured from 10-day-old chick embryos (EV-0 strain) by standard techniques and grown in DMEM supplemented with 8% FCS (Hyclone), 2% chicken serum (Life Technologies), 10% Tryptose Phosphate Broth, and antibiotics. Low-passage CEFs seeded into 100-mm dishes were transfected with 5 μg each of the desired recombinant retroviral vector DNAs by a standard calcium phosphate coprecipitation method (12). Cells were cultured 7 days, with passaging every other day, to promote propagation of the recombinant viruses. Complete propagation was confirmed by immunofluorescence staining with an antibody to the viral core protein p27 (SPAFAS, Inc.), and culture supernatant was harvested and stored at −80°C. Viral titers achieved in this manner were typically ∼106/ml. To generate CEF populations carrying two transgenes, cells were transfected with A-subtype viral vectors, cultured 4 days, and then infected with B-subtype viruses generated as above. Growth curves were determined by seeding 2 × 105 cells into 100-mm dishes and counting viable cells at various times later. Soft agar culture to assay anchorage-independent growth was performed by seeding 2.5 × 104 cells/well in triplicate into 6-well dishes, as described previously (13), except that concanavalin A was omitted. The BRK myc/p53ts cell line LTR.1A has been described previously (14). LTR.1A derivatives were generated using MSCV vectors as described (15) by selection in growth medium containing 0.5 mg/ml G418. Primary REFs (Whittaker Bioproducts) were transformed as described (16, 17).
The human Bin1-specific primers DN-sense (5′-AGT TCC CCG ACA TCA AGT CAC GCA-3′) and DN-antisense (5′-CTT GGC AAT TTT GGC TTC ATC C-3′), which span the 18-nt deletion in Bin1 DN, were used to document ectopic messages in cells. RT-PCR analysis was performed as follows. Total cytoplasmic RNA (2 μg) was mixed with 50 pmol of each primer, heated to 70°C for 5 min, and cooled rapidly on wet ice. The reverse transcription reaction was performed in 30 μl as suggested by the Mo-MLV reverse transcriptase vendor (Life Technologies). Ten percent of the reaction product was used as a template for 30 cycles of PCR (denaturation, 30 s at 94°C; annealing, 45 s at 55°C; polymerization, 60 s at 72°C). Ten percent of the PCR product was examined by agarose gel electrophoresis and photographed. For Northern analysis, 20 μg of total cytoplasmic RNA per lane were analyzed essentially as described (18). Blots were probed with cDNAs for murine ODC or murine CDC25A.
For Western blotting, cell lysates was prepared in NP40 buffer (Bin1 and Bcl-2) or RIPA buffer (c-Myc) using standard protocols. For c-Myc analysis, 1.5 mg of cell lysate were subjected to immunoprecipitation with 1 μg of anti-c-Myc antibody SC-42 (Santa Cruz Biotechnology) and 25 μl of protein G-agarose (Life Technologies); immunoprecipitates were washed three times with RIPA buffer before fractionation by nonreducing SDS-PAGE. For Bin1 and Bcl-2 analysis, 10 μg of cell lysate were fractionated directly by reducing SDS-PAGE. Gels were electrophoretically transferred to ECL membrane (Amersham) or Immobilon-P (Millipore) using standard methods (19). Blots were blocked in 3% skim milk and probed with the anti-Bin1 monoclonal antibody 99D (20), anti-c-Myc antibody 9E10, or anti-Bcl-2 antibody no. 124 (DAKO). Antibodies were diluted 1:50 in PBS with 2.5% skim milk and 0.1% Tween 20 and incubated with the membrane 12 h at 4°C. Blots were washed and incubated 1 h in the same buffer with secondary goat antimouse antibodies conjugated to horseradish peroxidase (BMB), and then developed using a chemiluminescence kit according to the protocol suggested by the vendor (Pierce). Indirect immunofluorescence was measured with anti-Bin1 99D as described (2).
CEFs (8 × 105) expressing the transgenes indicated were seeded overnight into 60-mm dishes in complete growth medium for apoptosis assays. Recombinant retrovirus-infected BRK/myc/p53ts cells were grown to ∼70% confluence in 100-mm dishes. At the start of the experiment, cells were washed twice with PBS and then placed in DMEM containing either 10 or 0.1% FCS. Thapsigargin (Calbiochem) was added to complete growth medium where indicated to a final concentration of 100 nm. At the end of the incubation period, cells were trypsinized, washed once with PBS, fixed, stained with propidium iodide, and processed for flow cytometry (10, 14). Alternatively, cells were harvested, stained with acridine orange (Sigma), and subjected to fluorescence microscopy to monitor for chromatin condensation. Staining with acridine orange was performed by dissolving 2.5 mg in 50 ml of PBS at room temperature, mixing 2 μl with 10 μl of cell suspension on a slide glass, and covering with melted paraffin wax.
Primary CEFs are a classical model system for assessing the biological functions of c-Myc. In these cells, c-Myc overexpression is sufficient to induce both transformation and apoptotic susceptibility. Thus, through a single genetic alteration, the effects of regulators or effectors of c-Myc function can be probed in a primary cell setting for transformation. Like rodent fibroblasts, CEFs undergo c-Myc-mediated apoptosis following serum deprivation, displaying characteristic cell detachment, blebbing, chromatin condensation, and DNA degradation (21). A major advantage of the CEF system is the availability of replication-competent retroviral vectors (7) that permit rapid and unselected gene transfer to a large population of cells, significantly reducing the selection for antiapoptotic background (a situation that develops rapidly in cells that constitutively express c-Myc). The role of Bin1 was investigated in this model by antisense and dominant inhibitory strategies using the RCOS BP or RCAS BP avian retrovirus vectors, which differ ∼10-fold in promoter strength (7).
Expression of Bin1 splice isoforms that interact with c-Myc was confirmed in CEFs by Western analysis and indirect immunofluorescence using the monoclonal antibody 99D (see below), which recognizes an epitope in the c-Myc-binding domain (20). Primary CEFs were infected with recombinant avian retroviruses that encoded either A or B envelope subtypes. An A-subtype vector was used to deliver human c-Myc or no insert, whereas a B-subtype vector was used to deliver no insert, one of two Bin1 dominant inhibitory genes tested, or Bcl-2 [the latter of which was used as a positive control for apoptosis suppression (22, 23)]. Cells were harvested several days after infection to verify transgene expression or to test in growth and apoptosis assays.
One dominant inhibitory mutant used (Bin1 DN) was a deletion mutant that lacked amino acids 143–148 (2), encoding KLVDYD, within the most evolutionarily conserved region of the NH2-terminal BAR domain. Deletions encompassing this region abolish antitransforming activity (8). Deletion of amino acids 143–148 similarly abolishes activity, but in addition, this generates a dominant inhibitory activity when assayed against wild-type Bin1 (see Fig. 1 A). This mutant had an intact MBD, so its dominant inhibitory activity could be mediated by interactions with c-Myc or with other proteins.
Transgene expression in cells infected with Bin1 DN vector was confirmed by RT-PCR (see Fig. 1,B). Effects on cell proliferation and neoplastic transformation were assayed by growth curve determination or growth in soft agar culture, which measures anchorage-independent capability. Bin1 DN did not significantly affect the ability of human c-Myc to drive CEF proliferation or neoplastic transformation as measured by these assays (see Fig. 1, C and D). Effects on apoptosis were assessed by serum deprivation. In the same manner as primary rodent fibroblasts (24) and consistent with earlier results (21), CEFs that overexpressed human c-Myc plus vector sequences underwent apoptosis within 24 h of serum deprivation, as defined by the appearance of terminal deoxynucleotidyl transferase (Tdt)-mediated nick end labeling-positive cells with sub-G1 phase DNA by flow cytometry and condensed chromatin by microscopic analysis. In contrast, under the same conditions, cells that coexpressed c-Myc plus Bin1 DN exhibited significantly less apoptosis. The reduction was specific insofar as Bin1 DN did not affect the efficiency by which apoptosis was activated by thapsigargin, a proapoptotic agent that acts by elevating intracellular calcium levels (see Fig. 1 E). Experiments with an antisense Bin1 gene that suppresses Bin1 function in mouse cells (25) yielded results in chick cells similar to those obtained with Bin1 DN (data not shown).
A role for c-Myc-Bin1 interaction in apoptosis was tested directly using a second dominant inhibitory strategy based on overexpression of the Bin1 MBD. Previous work demonstrated that the MBD is necessary and sufficient for interaction with c-Myc and that it has dominant inhibitory activity when overexpressed in cells (2). Expression of the transgenes and the endogenous c-Myc-binding isoform of Bin1 was confirmed by Western analysis (see Fig. 2,A, inset). In the absence of c-Myc, neither the MBD nor Bcl-2 significantly affected the proliferation rate. In c-Myc-expressing cells, Bcl-2 slightly retarded proliferation, whereas the MBD slightly promoted proliferation (see Fig. 2,A). As expected, c-Myc induced neoplastic transformation, but in the absence of c-Myc neither the MBD nor Bcl-2 could transform cells (see Fig. 2,B). Effects on apoptosis were assessed as before. In the absence of c-Myc, apoptosis was suppressed by Bcl-2 but not by the MBD. However, in the presence of c-Myc, both Bcl-2 and the MBD suppressed apoptosis (see Fig. 2,C). Bcl-2 also inhibited apoptosis induced by the calcium-mobilizing agent thapsigargin, consistent with previous reports (26), whereas this was not the case with the MBD (see Fig. 2 D). Thus, the suppressive effects of the MBD against c-Myc were specific. The effect of the MBD could be titered by elevating c-Myc expression through the use of the stronger RCAS BP vector (7), in support of the expectation that the MBD acts by effectively competing for c-Myc interaction (data not shown). We concluded that Bin1 participates in a mechanism used by c-Myc to induce apoptosis in transformed cells and that protein-protein interaction is part of the mechanism.
We observed no significant effect of full-length Bin1 vectors on the viability or proliferation of primary CEFs (data not shown). These observations were consistent with previous demonstrations that Bin1 lacks apoptotic activity in untransformed cells, including primary REFs, C2C12 mouse myoblasts, IMR90 human diploid fibroblasts, or normal human melanocytes (2, 3, 4, 25). In addition, we did not observe any augmentation of apoptosis when Bin1 was coexpressed with c-Myc, indicating that Bin1 was not limiting in CEFs under the experimental conditions tested. Taken together, we interpreted the findings to mean that Bin1 conditionally influences apoptotic susceptibility in a manner that depends on c-Myc and on neoplastic transformation.
We corroborated these results in a second primary cell system in which c-Myc drives apoptosis through a p53-independent mechanism(s). LTR.1A is a primary BRK epithelial cell line transformed by overexpression of c-Myc and a temperature-sensitive dominant inhibitory mutant of p53. In these cells, serum deprivation induces apoptosis at the nonpermissive temperature for wild-type p53 function (14). LTR.1A cells were infected with recombinant murine MSCV retroviruses, and the neomycin resistance cassette on each vector was selected for by culturing cells in G418. Transgene expression in drug-resistant cell populations was confirmed by RT-PCR analysis, and effects on proliferation and apoptosis were assessed as before. The MBD did not affect proliferation or apoptosis in LTR.1A cells, possibly because of the exceptionally high level of c-Myc expression in this system (14). In contrast, Bin1 DN promoted cell proliferation relative to the vector control (see Fig. 3,A), perhaps reflecting the different p53 status of LTR.1A (mutant) versus CEFs (wild-type). When p53-independent apoptosis was induced by depriving cells of serum and maintaining them at the nonpermissive temperature for p53 function (14), vector cells exhibited strong apoptotic death, but Bin1 DN significantly suppressed this effect (see Fig. 3, B and C). As before, experiments using an antisense Bin1 gene yielded results similar to those obtained with Bin1 DN (data not shown). The antiapoptotic effects were not explained by a reduction of c-Myc levels, ruling out the trivial possibility that the transgenes suppressed c-Myc expression (data not shown). We concluded that Bin1 participates in a mechanism of c-Myc-mediated apoptosis in transformed primary cells.
This study offers evidence of a role for the adapter protein Bin1 in cell death signaling in primary cells that are neoplastically transformed by c-Myc. This role was manifested in chick fibroblasts, where p53 is likely to be important to c-Myc activity, but also in primary epithelial cells, where p53 is dispensable (1). Therefore, Bin1 may mediate death or death sensitization signals that act independently but also cooperatively with p53 (see Fig. 4). Consistent with previous observations, Bin1 was not proapoptotic when overexpressed in primary cells by itself. This argues that Bin1 is not an “executioner,” but that it instead provides a conditional linkage between c-Myc and death effector machinery. Phosphorylation may be a germane condition insofar as both c-Myc and Bin1 are phosphoproteins (20, 27) and there is evidence of a specific apoptotic role for serine 71 in c-Myc (28), a site that is immediately proximal to the Myc Box 1 region implicated in Bin1 binding (2).
Our findings provide direct support for the “dual signal” model for c-Myc function based on interactions with a binding protein. How Bin1 acts remains to be determined. Overexpression of Bin1 in tumor cells engages a death process that is caspase independent (3) but reminiscent of the death process engaged by c-Myc in the presence of caspase inhibitors (29). Thus, c-Myc may activate multiple death processes, perhaps only a subset of which involves Bin1 or caspases. Although c-Myc can regulate transcription, no target gene has been associated unambiguously with biological function (30), and although Bin1 can regulate transactivation by c-Myc (8), we saw no changes in the expression of two c-Myc target genes, ornithine decarboxylase and CDC25A, that have been linked to c-Myc-mediated apoptosis in established cells (31, 32). Bin1 might also act via a transactivation-independent mechanism. The latter possibility can be entertained because (a) the structure of Bin1 suggests a signaling role rather than a transcriptional function, (b) not all biological actions of c-Myc can be ascribed strictly to gene activation, and (c) c-Myc-induced apoptosis may involve separate “priming” and “triggering” steps, based on the ability of c-Myc to trigger cell death when protein synthesis is inhibited (33).
Bin1 shares structural similarity outside the MBD with amphiphysin and yeast Rvs proteins, which have been implicated in endocytosis, actin regulation, and nuclear processes (34, 35, 36). Interestingly, however, deletion of the Bin1 gene in transformed mouse fibroblasts or in the fission yeast Schizosaccharomyces pombe results in defective apoptotic signaling or cell cycle control, respectively, but not in grossly altered cytoskeletal actin or endocytosis, suggesting that a full understanding of Bin1 function in animal cells has yet to emerge.5 In any case, the results of this study suggest one explanation for the frequent suppression or inactivation of Bin1 in human cancer cells (4, 5, 6): this event may help defeat the death penalty associated with c-Myc activation and in so doing help promote malignant progression.
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.
This work was supported in part by grants from the United States Army Breast Cancer Research Program (Grant DAMD17-96-1-6324), the Wistar Cancer Core (Grant CA10815), and the American Cancer Society (Grant CN-160).
The abbreviations used are: BRK, baby rat kidney; LTR, long terminal repeat; MBD, Myc-binding domain; REF, rat embryo fibroblast; CMV, cytomegalovirus; CEF, chick embryo fibroblast; RT-PCR, reverse transcription-PCR.
A. J. Muller, J. B. DuHadaway, P. S. Donover, J. F. Mostochuk, W. Du, E. Routhier, C. F. Albright, and G. C. Prendergast, unpublished results.
We are grateful to D. Beach, J. Cleveland, and R. Press for cDNA clones and to D. H. Crouch, G. E. Farmer, T. Halazonetis, and A. J. Muller for criticism. The support of A. I. Oliff and R. B. Stein and the assistance of the Wistar Flow Cytometry and Oligonucleotide Core Facilities is kindly acknowledged.