INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Oncogenic transformation is a multistep process that involves the loss of cellular growth controls maintained by tumor suppressor genes, as well as the acquisition of enhanced proliferative capacity and morphological alterations provided by dominant oncogenes (for reviews, see references 4 and 85). In support of this model, primary cells have been shown to be resistant to transformation by dominant oncogenes acting alone and seem to require the concerted action of at least two oncogenes to become transformed (23, 46-48, 71; for a review, see reference 37). Immortalized fibroblasts, on the other hand, are much more sensitive than similar primary cells to transformation by a variety of dominant oncogenes, including v-ras and v-src (21, 38, 55). The mechanism(s) behind this resistance to transformation is beginning to be understood and may include specific cell cycle blocks (30, 77, 82). In combination with the fact that more than 80% of human solid tumors are of epithelial origin, these observations suggest that primary epithelial cells would provide an excellent model system to study molecular signals involved in carcinogenesis.

The adenovirus (Ad) E1A gene is one of several DNA tumor virus oncogenes that have been used to study oncogene cooperative transformation of primary cells (for a review, see reference 2). E1A has been shown to immortalize primary cells, such that they maintain many of their differentiated characteristics (35, 64, 65). In addition, E1A can cooperate with Ha-ras and Ad E1B, as well as polyomavirus mT (pmT), to transform primary cells (14, 71, 81, 88). The pmT binds to and activates the c-Src protein (5, 10). This c-Src stimulation is required for cellular transformation by pmT (39). However, pmT cannot transform primary cells without the assistance of an immortalizing oncogene (11, 71). In the present study, we have extended these observations by investigating the ability of v-src to transform primary epithelial cells in the presence or absence of E1A expression. When baby rat kidney (BRK) cells were transfected with v-src, inefficient transformation was observed. However, when v-src was cotransfected with Ad 5 E1A 12S, marked improvement in transformation efficiency was noted. Synergistic enhancement of v-src transformation did not require full immortalization functions of E1A but only functions encoded by the conserved regions 1 and 2 (CR1 and CR2), corresponding to the ability of E1A to activate quiescent BRK cells into the cell cycle (65) and to the ability to bind the pRb family of proteins (16, 89). Transformation by Ha-ras is subject to dramatic modulation of transformation by E1A (13, 80). This was not observed with v-src cooperation. v-src-transformed cells exhibited a loss of actin organization, in the presence or absence of E1A, and inhibition of v-Src with radicicol (45) caused reestablishment of prominent actin stress fibers. This suggests that the transformation by E1A and v-src may be reversible and that the morphologic alterations observed were due to the kinase activity of v-Src and not to functions encoded by E1A. This is in contrast to transformation of primary epithelial cells by Ha-ras and E1A. These data imply that activation of the cell cycle in primary epithelial cells by E1A is sufficient to allow efficient oncogenic transformation of such cells by v-src and that this transformation is not subject to modulation by E1A.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Cells, transfections, and plasmids. 293 cells are human kidney cells transformed by Ad E1A and E1B (26) and were maintained in Dulbecco modified Eagle medium (DMEM)-5% fetal calf serum (FCS). BRK cells were prepared from 5-day-old rats as previously described (71). Transfections on BRK cells were performed at 2 days postplating by a modification of the calcium phosphate precipitation technique (90), as previously described (71). pJH v-src (expressing the wild-type [WT] SH2 and SH3 domains fused to the COOH-terminal portion of the Prague A v-src, which contains the kinase domain but lacks the regulatory tyrosine 527) was used at 5 µg/60-mm-diameter dish, as were the E1A expression plasmids which have been described previously (13). To visualize transformed foci, cells were fixed with methanol at 14 days posttransfection and stained with Giemsa (Sigma, St. Louis, Mo.). For cell growth curves and Western analysis, whole transfected plates (>20 foci/plate) were trypsinized to establish a polyclonal cell line and were used at low passage numbers (<10) in order to minimize the effects of clonal variation, except for v-src alone, which was derived from three separate clones that were pooled prior to assays. In addition, individual foci were cloned for additional growth curves and soft agar assays (Table 1). For cell growth curves, 10⁵ cells were seeded onto 60-mm-diameter dishes and grown in DMEM-5% FCS for the indicated number of days. Each time point represents at least four cell counts with a hemacytometer. The curves represent two independent experiments. Radicicol was used as previously described (58, 93), except that incubation in radicicol was carried out for 12 h (for Western blots) or 24 h (for immunofluorescence).

Western blot analysis. Western blot analysis was performed essentially as described previously (58, 93). Briefly, approximately equal numbers of cells were labeled with (approximately 1 µCi/2 × 10⁶ cells) Trans-³⁵S-Label (ICN, Costa Mesa, Calif.) overnight in DMEM with 5% FCS. The cells were washed twice with phosphate-buffered saline (PBS), lysed in boiling 2× Laemmli sample buffer, scraped into tubes, and boiled for an additional 5 min. Lysates were centrifuged to remove debris, normalized to each other by scintillation counting, and separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Proteins were transferred to Immobilon SP in a Bio-Rad Western blot apparatus. The filters were blocked overnight in blocking buffer (0.7% fish gelatin in 10 mM Tris [pH 7.5], 150 mM NaCl, 0.1% Tween 20 [TBST]). The blots were probed with either anti-c-Src antibody (SC-019; Santa Cruz Biotechnologies, Inc., Santa Cruz, Calif.) or an antiphosphotyrosine antibody (Transduction Laboratories, Lexington, Ky.) in blocking buffer, followed by three washes with TBST. Antibodies were detected with anti-mouse-horseradish peroxidase conjugate (Sigma), followed by a chemiluminescence detection method, essentially as previously described (58).

Immunofluorescence and photomicroscopy. Cells were seeded onto glass coverslips and maintained as described above. The cells were washed with cytoskeleton buffer (32) and then fixed with 3.7% paraformaldehyde for 10 min at room temperature, washed twice in cytoskeleton buffer, and permeabilized in 0.5% Triton X-100 in PBS. Nonspecific binding was blocked with 0.5% FCS-0.5% bovine serum albumin in PBS, and then cells were incubated in 1 µg of phalloidin-fluorescein isothiocyanate (FITC) per ml for 30 min, washed three times with PBS, and mounted in Airvol (Air Products, Allentown, Pa.). Fluorescence photomicrographs were taken with a Zeiss Axiophot epifluorescence microscope fitted with an Optronics DEI-750 video camera system. For phase photomicroscopy, an Olympus CK2 microscope fitted with Minolta X-700 was used with Kodak Tri-X Pan ASA 400 film.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

E1A first-exon functions enhance v-src transformation of BRK cells. To investigate possible cooperative transformation of primary epithelial cells by E1A and v-src, a plasmid expressing v-Src was transfected alone or cotransfected with a plasmid expressing either WT or mutated E1A 12S genes onto neonatal BRK cells. Five functional regions of E1A have been described, and mutations in all of these are represented by the E1A mutants shown in Fig. 1 and have been previously described (64, 88). As shown in Fig. 2, transformation of BRK cells by v-src alone was inefficient; the visible foci that were obtained were not observed until 18 to 28 days posttransfection (data not shown). This is in contrast to Ha-ras transfections, which never yield foci in the absence of E1A (71) (Fig. 2). Transfections of v-src onto NIH 3T3 cells resulted in efficient transformation (data not shown). When BRK cells were cotransfected with v-src and WT 12S, efficient transformation was observed, yielding numerous foci within 14 days posttransfection, similar to the transformation observed with Ha-ras (Fig. 2). All of the v-src-transformed cells tested had the ability to grow as soft agar colonies, including the rare cells transformed by v-src alone (Table 1). This cooperative transformation was not abrogated by mutations in the second exon of E1A, indicating that immortalization by E1A (which requires the second exon) is not necessary for v-src transformation, in contrast to E1A cooperation with Ad E1B (14, 15, 81). However, mutations that have been shown to interfere with E1A's ability to stimulate cell cycle progression in quiescent cells also interfered with v-src cotransformation. Specifically, deletion of the NH2 terminus and CR1 (NTdl814, containing amino acids 83 to 243 [88]) or deletion of CR2 (dl891-1339, lacking amino acids 110 to 174 [88]) of E1A abrogated cooperative transformation by v-src. Interestingly, pm563, which has been shown to be unable to cooperate with Ha-ras (88), was still able to cooperate with v-src (Fig. 2). pm563 does not form stable complexes with the transcriptional coactivator p300 (16, 89). However, this mutant does have the ability to stimulate cell cycle progression, although the proliferation cannot be maintained (67). Thus, the ability of E1A to synergize with v-src to transform primary epithelial cells is contained in the first exon and correlates with activation of the cell cycle by E1A.

View larger version (16K): [in this window] [in a new window]   FIG. 1.   Properties of E1A 12S mutants. The mutants have been described previously (64, 88) and are summarized here. Functional regions are indicated at the top. Protein binding indicates the ability of the protein to bind either Rb or p300 as well as WT E1A (+) or a level of binding that is less than that of WT (). Data for the columns headed immort. (immortalized) and activation are derived from references 67 and 64; data for pRb and p300 are derived from references 89 and 24; data for ras were derived from references 88 and 13; data for E1B were derived from reference 74; and data for v-src were derived from the present study. aa, amino acids.

View larger version (50K): [in this window] [in a new window]   FIG. 2.   Transformation of BRK cells by v-src and E1A mutants. Cells were transfected with plasmids expressing the various mutants of E1A and v-src, as described in Materials and Methods. The plates were fixed at 14 days posttransfection and stained with Giemsa and represent four or more dishes in at least two independent experiments.

Expression and kinase activity of v-src is not consistently altered by E1A. E1A has been shown to affect expression of a variety of cellular genes (for a review, see reference 2). To investigate whether E1A had an effect on the overexpression of v-src, cell lines were isolated from the transfections on BRK cells and Western blot analysis was performed (Fig. 3). No significant differences were observed in the expression of v-src in the presence of WT or mutated 12S. Thus, the enhanced transformation in the presence of E1A was not due to differential expression of v-src.

View larger version (44K): [in this window] [in a new window]   FIG. 3.   Analysis of v-src overexpression. Cells were cloned from the indicated transfections, and normalized lysates were processed for Western blotting and probed with an anti-Src antibody. 10 BRK, primary BRK cells.

View larger version (69K): [in this window] [in a new window]   FIG. 4.   Expression of v-src leads to increased phosphorylation of cellular proteins and is inhibited by radicicol. Total cell lysates were prepared from the cells indicated, which had been incubated in the presence (+ [lanes 5 to 7]) or absence ( [lanes 1 to 4 and 8 to 10]) of radicicol (rad; 0.1 µg/ml) for 12 h prior to lysate preparation. The blot was probed with antiphosphotyrosine antibody. Lanes 8 to 10, lower exposure of lanes 2 to 4.

E1A enhances the proliferative properties of v-src-transformed epithelial cells. As shown above, of the five functional regions of E1A, only the two regions required for cell cycle activation are required to cooperate with v-src. Therefore, the growth rates of BRK cells transformed with v-src or v-src plus E1A (WT or mutant) were analyzed. As shown in Fig. 5A, cells transformed by mutant or WT 12S and v-src had higher growth rates and saturation densities, compared to primary epithelial cells transformed by v-src alone. In normally cycling populations, the percentage of cells found in S phase was approximately 50% higher among BRK cells transformed by WT 12S and v-src, compared to v-src alone (Fig. 5B). In the case of dl891-1339 plus v-src, the resultant cells were not effectively established into a cell line, such that they failed to proliferate beyond early passages, which was exhibited in growth curves (Fig. 5A). The dl891-1339-plus-v-src-transformed cells appeared to experience a block in G₂/M and/or G₁ (Fig. 5B). This was not a clonal aberrance, since the experiment was done with a pool of multiple, independently arising foci. This may have been due to loss of Src expression in these cells, since Src has been shown to be required for G₂ progression (70). It has been previously shown that deletions of CR1 lead to a dominant negative effect on the establishment of primary cells by WT 12S (50). In a similar manner, dl891-1339 may not allow the stable establishment of primary cells, but this remains to be determined. Thus, the ability to activate the cell cycle not only is required for the initial transformation enhancement but also has a significant effect on the growth properties of v-src-expressing cells.

View larger version (34K): [in this window] [in a new window]   FIG. 5.   (A) Cell growth curves. Cells from the indicated transfections were seeded at 105 cells per dish, and growth rates were measured as described in Materials and Methods. Each data point represents an average of four counts. (B) FACS analysis of cells transformed with v-src, v-src plus dl891-1339, and WT 12S plus v-src. Cells were stained with propidium iodide and sorted as described in Materials and Methods.

E1A WT 12S is unable to modulate v-src transformation. Functions encoded by the second exon of E1A have been shown to down-modulate Ha-ras transformation of primary epithelial cells, such that they retain some of their epithelial morphology and characteristics (25). Deletion of the regions encoding these functions leads to a marked increase in transformation with ras (13, 80), but not E1B (14). This hypertransformation is concomitant with dramatic alterations in cell growth rates, adhesion, and actin cytoskeleton organization (20a). To assess whether v-src transformation was subject to such modulation imposed by E1A second-exon sequences, cell morphologies and growth rates among cells transformed with WT 12S plus v-src, CTdl976 plus v-src, and v-src alone were compared. As shown in Fig. 6, none of the v-src transformants maintained epithelial cell-style morphology or cell-cell junctions; instead, they exhibited a star-shaped morphology, even in the presence of WT 12S. Deletion of some or all of the second exon did result in cells that were somewhat smaller and more refractile and enabled them to grow to higher saturation densities (Fig. 5A), which is similar to what has been observed in E1A cooperation with Ha-ras.

View larger version (136K): [in this window] [in a new window]   FIG. 6.   Cell morphologies of transformed or immortalized (imm) BRK cells. Photomicrographs were taken at equal magnifications. Bar, 100 µm; 10 BRK, primary BRK cells.

v-src transformation of primary epithelial cells leads to a loss of actin stress fibers, which are maintained in E1A-immortalized BRK cells. c-Src has been shown to be involved in the assembly of focal adhesion complexes and activation of integrin-mediated signals, primarily due to its ability to form stable complexes with FAK (40, 74, 75). c-Src kinase activity is involved in the regulation of epidermal growth factor (EGF)-dependent, actin cytoskeleton rearrangement via phosphorylation of rhoGAP p190 (7). v-Src, on the other hand, has been demonstrated to cause disassembly of actin stress fibers and focal adhesions in fibroblasts (20, 42). Similarly, other viral oncoproteins have been observed to affect actin stress fibers (59, 86). E1A has been shown to induce and maintain epithelial cell differentiation and morphology (Fig. 7C) (8, 22, 25, 64, 66). Therefore, the organization of F actin in src-plus-E1A-transformed versus E1A-immortalized BRK cells was analyzed.

View larger version (141K): [in this window] [in a new window]   FIG. 7.   F-actin organization of E1A-immortalized epithelial cells versus E1A-plus-v-src-transformed epithelial cells. Cells were stained with phalloidin-FITC, as described in Materials and Methods. (A and B) WT 12S-plus-v-src-transformed cells; (C) WT 12S-immortalized cells; (D and E) HBdl12-plus-v-src-transformed cells; (F) HBdl12-immortalized cells; (G and H) v-src-transformed cells; (I) primary BRK cells; (B, E, and H) cells exposed to 0.1 µg of radicicol per ml for 24 h prior to fixation. Bar, 30 µm.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The results presented here show that transformation of mammalian primary epithelial cells by v-src alone is inefficient, in contrast to primary fibroblasts (34) or established fibroblast lines such as NIH 3T3 cells. However, coexpression of E1A leads to efficient transformation of these cells, such that they are able to grow indefinitely and in an anchorage-independent fashion. This synergistic cotransformation ability mapped to the sequences contained in the first exon, in particular, CR1 and CR2 of E1A 12S. Interestingly, a deletion mutant lacking CR2 (dl891-1339) produced few transformed foci, which did not produce stable cell lines. Furthermore, an NH₂-terminal deletion mutant lacking CR1 (NTdl814) did not produce transformed foci with v-src. That the latter did not enable src transformation was not surprising, since our lab has previously found NTdl814 to actually inhibit WT 12S functions (50). These observations underscore the requirement for CR1 and CR2 functions for v-src transformation. An NH₂-terminal point mutant, pm563, which lacks the ability to bind p300 (89), was able to enhance v-src transformation, indicating that the requirements for v-src cooperation are less stringent than those for E1B or Ha-ras cotransformation (14, 79, 88). Indeed, neither E1B, Ha-ras, nor pmT, which activates c-src (5, 10, 39), is able to transform primary cells alone, but each requires multiple but different functions encoded by E1A (14, 71, 79, 81, 88). Nevertheless, v-src alone is able to transform primary epithelial kidney cells, albeit inefficiently. Previous studies have shown that primary rodent (but not human) fibroblasts can be transformed by v-src alone (34). Susceptibility to transformation by v-src in primary cells appears to be partly dependent on cellular senescence and aging (82). Also, it has been shown that adrenocortical cells can be inefficiently transformed by v-src and that v-myc can cooperate with v-src to efficiently transform such cells (48). The inefficient transformation by v-src alone in this study as well as previous studies could rely on the acquisition of mutations in cell cycle regulatory genes potentially reproducing the effects of E1A expression. While this hypothesis remains unconfirmed, heterogeneity in primary cultures may be relevant to the observation of rare v-src transformants (83). It seems likely that the inefficiency of transformation observed here may be due at least in part to the fact that BRK epithelial cells undergo senescence and apoptosis after 24 h in culture, which E1A is able to rescue (63). In addition, the ability of E1A, via CR1- and CR2-mediated binding of the pRB family (6, 89), to activate the cell cycle probably enables the somewhat higher growth rates observed in the E1A- and src-transformed cells, compared with the cells transformed by src alone. The ability of E1A to activate quiescent cells into the cell cycle seems crucial for the enhancement of v-src transformation of primary epithelial cells.

Hyperphosphorylation of cellular proteins by v-Src seems to be enhanced by WT 12S, but not by mutants lacking the second exon, such as CTdl976. The significance of this is not clear at present, since CTdl976 is as competent in transformation assays with v-src as WT 12S. Furthermore, the enhanced kinase activity does not correlate with substrate-independent growth or growth rate in vitro (Fig. 5; Table 1). Other investigators have observed inconsistent increases in tyrosine phosphorylation by src in cells expressing E1A (1), but the data neither indicate the significance of nor suggest a mechanism for the increase. We find here that the mechanism does not involve increased expression of the v-src oncogene. However, differences in localization, stability, or specific activity have not been established at present.

Requirements for transformation of primary epithelial cells by v-src and E1A are different from those of other cooperating oncogenes (Fig. 1). For efficient transformation of primary epithelial cells by v-src and E1A, the ability of E1A to induce cell cycle via CR1 and CR2 is apparently the only requirement. This is interesting, because previous studies have suggested that immortalization of primary cells might be required for efficient v-src transformation (38, 48), while we have observed here that mutants incapable of immortalization by themselves are able to cooperate with v-src to transform primary epithelial cells (Fig. 1). Thus, E1A mutants lacking the ability to establish cell lines alone can cooperate with v-src to establish transformed lines. In fact, the requirements for E1A to cooperate with v-src in transforming primary epithelial cells are less stringent than those for Ad E1B or even Ha-ras. This is further reflected in the ability of v-src to transform such primary cells alone, albeit inefficiently, while neither Ad E1B nor Ha-ras is able to do so (47, 71).

The role of p300 binding by E1A in cell transformation has been found to involve changes in gene expression (for reviews, see references 2 and 54). Similarly, changes in cellular gene expression can also be induced by v-src transformation (27, 52, 76). Furthermore, it has been shown that p300 binding by E1A can contribute to cell proliferation (36, 92), by interfering with the expression of p21^WAF (3, 12, 53, 87). p27^KIP and p21^WAF are cyclin-dependent kinase inhibitors with many targets in common (9, 17, 19, 31, 41, 56, 60, 61, 84). However, p27^KIP can be inhibited by E1A directly, apparently independently of transcription alterations (49). It is possible that primary kidney epithelial cells (as opposed to the PC12 cells used in the aforementioned studies) are prevented from proliferation primarily by p27^KIP and not p21^WAF, although this has not been formally tested. Since p300 binding does not seem to be required for E1A's cooperative transformation with v-src but is required for ras cotransformation (for a review, see reference 2), it seems possible that v-src may allow cells to escape a proliferation block by a p21^WAF-dependent mechanism. Given that p27^KIP and p21^WAF have been shown to be involved in attachment-dependent growth (19) and that inhibition of Src kinase activity prevents soft agar growth (Table 1), it is possible that Src kinase activity allows cells to circumvent such cell cycle regulation steps, as has been suggested elsewhere (72).

The data presented here also show that while E1A is able to enhance transformation efficiency, v-src-mediated morphological alterations are not affected dramatically by E1A. These alterations include loss of cell-cell contact, actin stress fibers and/or microfilaments, and focal adhesion plaques. Radicicol, a potent and reversible inhibitor of Src kinase activity (44, 45, 58), is able to reverse the loss of actin stress fibers induced by v-Src, although cells do not regain cell-cell contacts. Although radicicol has been shown to be specific to Src kinase (45), it can inhibit mos and ras transformation as well (93). The fact that v-src-transformed primary epithelial cells exhibit loss of stress fibers is hardly surprising, given the wealth of data showing this phenomenon in other cell types (for a review, see reference 57 and references therein). However, c-Src has a positive effect on cell spreading and focal adhesion formation on fibronectin, but this role of c-Src does not seem to require its kinase activity (40). However, it is interesting that E1A is not able to suppress v-src-mediated morphological transformation of epithelial cells, given E1A's ability to do so with an activated ras gene (13, 20a, 25).

The ability of v-src alone (present study and reference 34) but not Ha-ras (present study and references 47 and 71) to transform primary cells suggests further differences between the signals mediated by ras and src. Both ras and src activation cause stimulation of the ERK-MAP kinase cascade via c-raf activation (18, 51, 91). The mechanisms employed by Src and Ras to accomplish this seem to be different (78). The activation of raf, however, does not lead to full transformation of either NIH 3T3 cells (43, 62) or primary epithelial cells, even with E1A (20a), and activation of rho-related proteins is required in combination with the activation of raf for full transformation. The activation of Rho family proteins by ras is well documented (for a review, see reference 68). src is also able to transduce signals through the Rho family G proteins via their GAPs (7, 33). However, activated ras and activated src have been shown to have opposite effects on actin stress fibers in fibroblast cells, presumably by their actions on Rho activity (7, 69). In BRK cells, however, cotransformation by either Ha-ras or v-src leads to a loss of stress fibers, again suggesting common cellular effects for these oncogenes, although the mechanisms may be different. Perhaps the differences in their abilities to transform primary epithelial cells alone, or in conjunction with E1A, can be explained by the observation that src can act upstream of ras, as some signal transduction cascade models would suggest (73). This might indicate that the activities or signals induced by activated ras are but a subset of those induced by activated src in epithelial cells. In addition, since Src mediates important alterations in epithelial cell transformation directly by its kinase activity (29), it may be a more potent oncogene than ras in epithelium-derived cells (48).