Growth-regulated cells, such as 3T3 mouse embryo fibroblasts (MEFs), require more than one growth factor for growth, usually the insulin-like growth factor I (IGF-I) in combination with either platelet-derived growth factor or epidermal growth factor. Singly, these growth factors cannot sustain the growth of 3T3 cells. However, if the IGF-I receptor (IGF-IR) is even modestly overexpressed, then IGF-I, by itself, stimulates the growth of MEFs in monolayer and makes them capable of forming colonies in soft agar. The granulin/epithelin precursor (GEP) has been identified as the only growth factor, thus far, that can stimulate by itself the growth of R− cells, a 3T3-like cell line in which the genes for the IGF-IR have been deleted. We have expressed GEP in R− cells and show that these cells can now grow in serum-free medium. GEP, however, cannot replace other functions of the IGF-IR, such as protection from apoptosis (anoikis) or transforming activity (colony formation in soft agar). GEP activates, in R− cells, the two signaling pathways that are known to be sufficient for IGF-I-mediated mitogenesis in cells overexpressing the IGF-IR, the mitogen-activated protein kinase and the phosphatidylinositol 3-kinase pathways. This may explain why GEP, by itself, can replace the IGF-IR for growth in monolayer cultures. It also confirms that, for transformation, other pathways must be activated besides the two pathways that are sufficient for mitogenesis.
Although many cell types (especially tumor cells) can grow in SFM3 because they usually produce their own growth factors, growth-regulated cells, like 3T3 MEFs or human diploid fibroblasts, require at least two growth factors for growth in monolayer cultures (1, 2, 3). In general, a combination of IGF-I and either EGF or PDGF are sufficient to sustain the growth of ordinary 3T3 cells or other growth-regulated cells. When the IGF-IR is overexpressed, even modestly, IGF-I (or IGF-II) induces mitogenesis in the absence of other growth factors (4, 5). Thus, 3T3 cells with 15 × 103 IGF-IR/cell still require PDGF and IGF-I for growth, but an increase to 22 × 103 receptors/cell is already sufficient to make the cells exclusively dependent on IGF-I (6). Under these conditions, it is generally believed that the IGF-IR, to induce mitogenesis, activates two different pathways (7, 8). The first pathway depends largely on IRS-1 (9, 10, 11, 12, 13) and goes through PI3-kinase, Akt/PKB, and p70S6K (11, 14, 15). The second pathway, mostly but not wholly dependent on Shc proteins (16), goes through Grb2, Sos, Ras, Raf, and MAPK (7). The Grb2 pathway can also be activated through IRS-1, which is known to have a binding site for Grb2 (17). The IGF-IR has other activities, besides its mitogenicity. It protects cells from apoptosis (18), and it can induce differentiation in certain types of cells (19, 20, 21) and cause transformation of MEFs as well as of other types of cells (18). A role of the IGF-IR in transformation is also substantiated by the finding that R− cells (devoid of IGF-IR) are refractory to transformation by a number of viral and cellular oncogenes (18), although v-src (but not an activated c-src) can transform them (22).
The GEP and the epithelins that derive from it are growth factors that have been known for some time (23, 24, 25, 26, 27) but the function of which has not been studied in detail. They are expressed by granulocytes (25), by several animal tissues (28), and by transformed cell lines (29, 30), and they bind to a putative receptor (31, 32), which has not yet been isolated or cloned. One interesting feature of GEP is that it stimulates the growth of R− cells (33, 34), which are 3T3 cells originating from mouse embryos with a targeted disruption of the IGF-IR genes (35, 36). R− cells grow in 10% serum but cannot be stimulated to grow by any other purified growth factor, including EGF and PDGF (30, 34, 37), which, in combination with IGF-I, stimulate the growth of 3T3 cells with endogenous IGF-IRs (1, 2). GEP is presently the only purified growth factor that can induce DNA synthesis and increase in cell number in R− cells in the absence of other growth factors (38).
Because the receptor for GEP (or its epithelins) has not yet been cloned, little is known about its signaling pathways and its functions. However, Zhang and Serrero (39) have reported recently that an antisense cDNA to the GEP sequence can inhibit the tumorigenicity of a teratoma PC cell line. Less relevant to the present investigation, but also of interest, is a report that the epithelins act as cofactors for the Tat proteins of the HIV (40).
Because GEP can stimulate growth in cells devoid of IGF-IR, we have asked whether GEP can replace the other functions of the IGF-IR, specifically protection from apoptosis and transformation. We also would like to know whether the mitogenicity of GEP in R− cells can be explained by an ability to activate the same pathways that the overexpressed IGF-IR activates for mitogenesis. For this purpose, R− cells are the cells of choice. They are very strictly growth regulated (22, 33), and they are refractory to growth stimulation by either PDGF (41) or EGF (37), the receptors of which share some transducing signals with the IGF-IR (see below). In addition, the absence of an IGF-IR excludes any possible interaction of GEP with the receptor itself. The two questions we have specifically asked, therefore, are: (a) to what extent can GEP replace the functions of the IGF-IR? and (b) is its mitogenicity in 3T3-like cells based on similar or different pathways? To answer these questions, we have stably transfected R− cells into a full-length GEP cDNA (25) and examined its effect on the growth characteristics and the transducing signals of these cells.
MATERIALS AND METHODS
The expression vectors containing the full-length human granulin precursor cDNA originated in the laboratory of Andrew Bateman (25). The vector pcEXV-3 is suitable for transient transfection; the vector pRc/RSV, carrying a neomycin-selectable marker, was used to stably transfect R− and BALB/c3T3 cell lines.
Plasmid pGR123 contains the human epithelin precursor fused in-frame to a Flag-tag (Kodak Flag Systems). This construct was created by the following procedure. The GEP cassette was excised by EcoRI digestion from the pcEXV-3 and ligated into the pcDNA1.1 (Invitrogen) EcoRI site dephosphorylated previously. This vector was then cut with NarI-XbaI, where NarI cuts in position 1680 of epithelin DNA and XbaI cuts in the polylinker of pcDNA1.1, 3′ of the epithelin insert. The epithelin-flag fusion was created by PCR using a direct primer (5′-TCACGTGGGTGTGAAGGACGTGGAG-3′) designed in the 5′ region of the NarI site in the epithelin cDNA and a reverse primer (5′-TGCATGCCTTCTAGAGAATTCTCACTTGTCATCGTCGTCC-TTGTAGTCCATCAGCAGCTGTCTCAAGGCTGGGTCCCTCAAAGGGG-CGTC-3′) containing the excised part of the epithelin sequence, the Flag tag and the XbaI site. Template for the reaction was the epithelin precursor cDNA excised with EcoRI from the pcEXV-3 vector. The PCR fragment was then digested with NarI-XbaI and ligated into the pcDNA1.1/epithelin vector cut previously with NarI-XbaI. The cDNA fragment containing the epithelin fused to a Flag-tag was then excised from the pcDNA1.1 vector with EcoRI and then cloned into the EcoRI site of the MSCV.pac vector, carrying resistance to puromycin and suitable for retroviral transductions. The retroviral vector MSCV.pac was provided by Dr. R. G. Hawley (University of Toronto, Toronto, Ontario, Canada) and is described elsewhere (42).
Plasmids pHIT60 and pHIT123 were gifts of Dr. A. Kingsman (University of Oxford, Oxford, United Kingdom) and are described elsewhere (43). pHIT60 contains the MLV gag-pol cassette under the control of the human cytomegalovirus immediate early (hCMVi.e.) promoter, whereas pHIT123 contains the hCMVi.e.-driven MLV ecotropic envelope. Both plasmids carry the SV40 origin of replication in their backbone.
R− and BALB/c3T3 cells were used in these experiments. BALB/c3T3 cells are mouse embryo cells that have been grown in our laboratory for several years. R− cells (34) are 3T3-like cells derived from mouse embryos with a targeted disruption of the IGF-IR genes (35, 36). They grow in 10% FBS, but they do not grow in SFM. The establishment of cell lines expressing GEP was as follows. R− cells were cotransfected with pRc/RSV and pPDV3+ (44), which encodes the puromycin resistance gene, and selected with 3 μg/ml puromycin. BALB/c3T3 were transfected with pRc/RSV and selected with 800 μg/ml neomycin. All transfections were performed using the Profection kit (Promega).
R− cells expressing the epithelin precursor and Flag were established by retroviral transduction. Transient calcium phosphate DNA transfections of 293T cells were carried out as described (45). MLV retroviral vectors were harvested and used for transducing target cells as described elsewhere (43, 46). Transducted cells were grown in DMEM supplemented with 10% FBS, 2 mm l-glutamine, and 2.5 μg/ml puromycin.
The fetal human kidney carcinoma 293T cell line was produced in the laboratory of Dr. David Baltimore and was purchased from American Type Culture Collection upon authorization of Rockefeller University. 293T cells were grown in DMEM supplemented with 10% FBS and 2 mm l-glutamine.
Northern Blot Analysis.
Total RNA was extracted using a Qiagen kit (Rneasy Mini kit), and RNA blots were carried out by standard procedures. Radioactive probe was prepared by the Random Primed DNA labeling kit (Boehringer Mannheim) and [α-32P]dCTP (∼3000 Ci/mmol; Amersham, Arlington Heights, IL). The 2.1-kb human epithelin fragment used as a probe was prepared by EcoRI digestion of the pcEXV-3 plasmid.
Epithelin/Flag Detection by Immunoprecipitation.
Semiconfluent cells were serum starved for 1 h using a SFM (DMEM; Life Technologies, Inc.) depleted of l-cysteine. Subsequently, cells were labeled in vivo with a [35S]cysteine (300 μCi/ml) containing medium for 3 h. Lysates from cells were immunoprecipitated overnight with anti-Flag mAb m2 (Kodak) plus protein G + protein A agarose conjugated. Proteins were separated on a 4–15% precasted acrylamide gel (Bio-Rad). The gel was fixed in 10% glacial acetic acid + 30% methanol and treated with “En3hance” (DuPont) before autoradiographic exposure.
Preparation of Conditioned Medium.
R− transfected cells were plated in 10-cm Petri dishes in 10 ml of DMEM containing 10% FBS and 3 μg/ml puromycin and were incubated at 37°C in a humidified CO2 incubator. After reaching subconfluence, the cells were washed three times in Hank’s solution and then incubated in 10 ml of SFM. The next day, this medium was discarded, and the cells were washed with Hank’s solution, again three times, and then fresh SFM was added. The medium from these cultures was collected after 4 days. Collected medium was sterilized using a 0.2-μm pore filter and stored frozen at −20°C until needed.
Growth in Monolayer.
All cell lines were passaged in DMEM containing 10% FBS. R− epi clones were cultured in DMEM containing 10% FBS and 3 μg/ml puromycin. To make cells quiescent, they were seeded at a density of 4 × 104 cells/plate in the presence of serum. After 24 h, the cells were washed with Hank’s solution and incubated in SFM [DMEM supplemented with 50 μg/ml transferrin (Sigma, St. Louis, MO) and 0.1% BSA (Sigma)]. The cells were left 48–72 h in SFM, which was renewed with fresh medium every 24 h, without additions. When we tested the cells for growth response to epithelin CM, the cells were stimulated with CM collected previously. Cell number was evaluated at 48 h after CM stimulation.
This parameter was determined by using a BrdUrd detection and labeling kit (Boehringer Mannheim, Indianapolis, IN). Briefly, to determine the cumulative index for DNA synthesis, quiescent cells (5 days in SFM) were incubated for the last 24 h in the presence of 10 μm BrdUrd. All cells were subsequently fixed in a solution containing 94.5% ethanol, 5% acetic acid, and 0.5% Triton x-100 (Sigma) and washed with PBS prior to immunostaining. Primary (mouse monoclonal BMC 9318) and secondary antibodies (sheep antimouse IgG FITC labeled) were applied at a dilution of 1:10 in PBS. To visualize all nuclei, DNA was additionally stained with 500 μg/ml bisbenzimide H33258 (Sigma). Vectashield (Vector Laboratories, Burlingame, CA) mounting medium was subsequently applied, and the BrdUrd labeling index was determined using a Zeiss microscope working in epifluorescence mode (×500). In randomly selected fields, 200 cells/dish were counted.
Cell Survival in Suspension.
To determine this parameter, quiescent cells were detached from a culture dish with 0.02% EDTA (disodium EDTA) and seeded on dishes coated with polyHEMA (Aldrich, Milwaukee, WI) prepared according to the methodology described previously (47). Cells were seeded in SFM at a concentration of 4 × 104 cells/dish and were left untreated. Twenty-four h later, cell suspensions were collected and dissociated with 0.25% trypsin. Viable cells (by trypan blue exclusion) were counted with a hemocytometer.
To assess the ability of epithelin-expressing clones in repairing wounds, the procedure of Dulbecco (48) was adopted. R− epi cells were plated in 35-mm dishes and cultured in DMEM containing 10% FBS and 3 μg/ml puromycin. After reaching subconfluence, the cells were washed three times with Hank’s solution and then incubated in 10 ml of SFM. After 48 h, a wound was created by scraping the monolayer, and the cells were incubated with 10 μm BrdUrd. After 48 h, the cells were stained for BrdUrd as described above.
Anchorage-independent growth was determined by a soft agar assay as described previously (33, 34). Cells (1 × 103) in growth medium (DMEM supplemented with 10% FBS) containing 0.2% agarose (Difco) were plated in 35-mm dishes with growth medium containing 0.4% agarose underlie. Cells were allowed to grow in soft agar for 2 weeks. Anchorage-independent growth was assessed by scoring the number of colonies >125 μm.
Cell lysates (500 μg of protein) were immunoprecipitated with an anti-Shc mAb (Transduction Laboratories) and a protein G + protein A agarose conjugated (Calbiochem) in the presence of HNTG buffer [20 mm HEPES (pH 7.5), 150 mm NaCl, 0.1% Triton X-100, 10% glycerol, 0.2 mm sodium orthovanadate, 0.2 mm phenylmethylsulfonyl fluoride, and 2 mg/ml aprotinin]. The immunoprecipitates were resolved on a 4–15% gradient gel by SDS-PAGE and transferred to a nitrocellulose filter. For immunoblotting, membranes were blocked with 5% dry nonfat milk in TBST buffer [10 mm Tris (pH 8.0), 150 mm NaCl, and 0.1% Tween 20] overnight at 4°C. To detect phosphorylation of immunoprecipitated proteins, we used an anti-phosphotyrosine horseradish peroxidase-conjugated antibody (PY20; Transduction Laboratories). Blots were developed with the ECL system (Amersham Corp.) according to the manufacturer’s instructions.
After stripping, the amount of the immunoprecipitated protein was detected by incubation with an anti-Shc mAb (Transduction Laboratories), followed by horseradish peroxidase-conjugated anti-mouse IgG. Grb2 was immunoblotted with a monoclonal anti-Grb2 antibody (Transduction Laboratories).
Cell lysates were collected directly from cultures of transduced cells growing in DMEM containing 10% FBS and 2.5 μg/ml puromycin. Semiconfluent cell dishes were lysed on ice with 300 μg/ml of lysis buffer [50 mm HEPES (pH 7.5), 150 mm NaCl, 1.5 mm MgCl2, 1 mm EGTA, 10% glycerol, 1% Triton X-100, 1% phenylmethylsulfonyl fluoride, 0.2 mm sodium orthovanadate, and 1% aprotinin]. Protein concentration was determined with a Bio-Rad protein assay (Bio-Rad, Hercules, CA), and 40 μg of proteins were separated on a 4–15% gradient SDS-PAGE (Bio-Rad) and transferred into nitrocellulose membranes. Blots were blocked with 5% nonfat dry milk in TBST [10 mm Tris-HCl (pH 7.5), 150 mm NaCl, and 0.1 Tween 20]. Secondary antibodies used were antimouse IgG or antirabbit IgG conjugated to horseradish peroxidase (Santa Cruz Biotechnology), which were subsequently visualized by a ECL detection reagents (Amersham).
Western blots of MAPK were performed on 10% precasted gels (Bio-Rad). Phosphorylation was detected using an Anti Active Mapk Ab from Promega. The whole amount of protein was estimated using a rabbit polyclonal antibody against Erk-1 from Santa Cruz Biotechnology. MAPK activation was also measured by densitometry, correcting the values for the amounts of proteins in each gel.
Akt phosphorylation and amounts were detected by using a Phosphoplus Akt (Ser473) Antibody kit from New England Biolabs. p70S6K phosphorylation was detected by using a Phosphoplus p70S6K Ab kit from New England Biolabs. This is a rabbit polyclonal antibody against phosphorylated thr421/ser424 of p70S6K. The protein amounts were detected by using p70s6k antibody (rabbit polyclonal IgG) from the same kit. For cyclin B1, we used the antibody Clone H-433 from Santa Cruz Biotechnology.
Establishment of Cell Lines Expressing the GEP.
In previous reports (30, 38), one of our laboratories had shown that purified GEP can stimulate the growth of R− cells, which are 3T3-like cells with deleted IGF-IR genes (33, 34). In the present experiments, we examined the effects of a transfected GEP cDNA (25) on the growth characteristics and signaling pathways in R− cells. R− cells were cotransfected with a plasmid expressing GEP and a second plasmid expressing puromycin resistance (see “Materials and Methods”). A number of clones were selected and tested for expression of GEP mRNA by Northern blots. Several clones tested positive for GEP mRNA (Fig. 1 A), although the intensity of expression varied from one clone to another. Most of them contained clones expressing a mRNA of the same size as the GEP mRNA of BRL-3A cells, a strong producer of GEP (30, 38). Some of the clones also expressed other mRNAs of different sizes. This is expected from clones of cells transfected with expression plasmids, which often are integrated in tandem with the selectable marker and give rise to products of read-through. Several clones were selected for further studies.
Expression of the GEP in R− Cells.
To facilitate the demonstration of GEP expression, we generated a plasmid with a Flag-tag (see “Materials and Methods”) and cloned it into a retroviral vector, MSCV.pac, which also carries the marker for puromycin resistance. R− cells were transduced with this retroviral vector or with the same vector carrying the puromycin resistance gene but no GEP sequence. For detection of GEP, the transduced cells were labeled with [35S]cysteine, and the CM was immunoprecipitated with a FLAG antibody, and the gel was autoradiographed (Fig. 1 B). The GEP product is clearly visible as a protein that travels slower (Mr 97,000) than expected (Mr 60,000), a not infrequent occurrence in gels. The FLAG epitope adds an additional eight amino acids, obviously not enough for an increase in molecular weight of that magnitude. We do not know whether this shift in mobility may be due to the presence of numerous cysteines in the GEP sequence or to posttranslational modifications, or both. The bands visible in both lanes above the FLAG-GEP protein are aspecific, as far as we could determine. The clones expressing the FLAG-tagged GEP showed similar growth characteristics to the clones transfected previously with the original GEP cDNA (see also below).
R− Cells Expressing GEP Grow in SFM.
Our interest in GEP originated from our observation that, thus far, it is the only purified growth factor that will make R− cells grow in SFM (38). Because the IGF-IR, when overexpressed, can make cells grow in IGF-I only, protect them from apoptosis (see the “Introduction”), and transform them (see also the “Introduction”), we asked whether expression of GEP would mimic the effects of an activated IGF-IR.
R− cells, like most 3T3 cells, do not grow at all in SFM, and indeed, they do not grow even when the SFM is supplemented with purified growth factors (30, 33, 34, 38). They grow only in 10% serum. R− cells expressing GEP can grow in SFM (Fig. 2), the extent of growth varying from one clone to another (only two clones are given here; additional clones will be examined in Fig. 3). The growth-promoting effect of the GEP cDNA on R− cells was reproducible. The growth of parental R− cells or R− cells transfected with the empty vector (R-Vec2) are given for comparison.
Because growth (cell number) is often a delicate measurement, we are also giving here BrdUrd incorporation into R− cells and R−/GEP cells after 5 days in SFM (Fig. 3 A). The cells were labeled with BrdUrd for the last 24 h. Clones expressing GEP incorporate BrdUrd in a high percentage, even after 4 days in SFM, whereas only a small fraction of R− cells or R− cells transfected with the empty vector were still synthesizing DNA at this time. These growth experiments have been repeated several times with the same results.
The GEP Does Not Protect R− Cells from Apoptosis.
The next question was directed at the ability of the IGF-IR to protect cells from apoptosis, specifically a form of apoptosis called anoikis (49, 50), that results when cells are denied attachment to a substratum (51, 52, 53). For this purpose, R−/GEP cells were seeded in polyHEMA plates, where R− cells die, but where R− cells expressing a sufficient number of IGF-IR can be rescued by IGF-I (49, 50). Fig. 3 B shows that R−/GEP cells die in polyHEMA plates as rapidly as the parental cells or the R− cells transfected with the empty vector. GEP, therefore, in the absence of the IGF-IR, cannot protect cells from anoikis, i.e., cannot substitute the IGF-IR under these conditions.
Because the GEP was discovered originally in granulocytes (25, 54) and is present in hematopoietic tissues (55), it was thought to have a function in wound healing. To test this function, we used the in vitro wound healing test described by Dulbecco (48) and detailed in “Materials and Methods.” Both R− cells and GEP-expressing cells were used. The cells were plated on coverslips and then made quiescent by serum deprivation. After wounding, the cells were incubated in SFM, labeled with BrdUrd, and examined 24 h later. Fig. 4 shows that R− cells have not filled the wound (Fig. 4,B), only a few cells are visible in the wounded area, and only one of them is labeled by BrdUrd (Fig. 4,A), as expected. R−/GEP cells, instead, have proliferated actively, have filled almost completely the wound, and are labeled with BrdUrd (Fig. 4, C and D).
Effect of GEP on Growth Characteristics of BALB/c 3T3 Cells.
Another characteristic of the IGF-IR is its ability to transform cells (56). Because R− cells are refractory to transformation (18), we tested GEP in both R− and BALB/c 3T3 cells, which can be transformed by a variety of oncogenes. For this purpose, R− or BALB/c 3T3 cells were stably transfected with the GEP plasmid. The mRNA expression levels for BALB/c 3T3 cells are shown in Fig. 5,A. The selected clones were then tested for growth in monolayer cultures. The expression of GEP causes these cells to enter DNA synthesis in SFM (not shown). The BALB/c and the R− clones expressing GEP were then tested in soft agar (Fig. 5 B). Neither R−/GEP nor BALB/c GEP clones form colonies in soft agar. As control cells, we used p6 cells. These cells, also derived from BALB/c 3T3 cells, express roughly 5 × 105 IGF-IRs/cell and are transformed (5), as expected from cells expressing such high numbers of IGF-IRs (4, 5). Indeed, p6 cells, which have been very stable over a period of almost 10 years, form many colonies in soft agar. It seems, therefore, that overexpression of GEP is not sufficient for transformation of 3T3 cells, whether or not they have endogenous IGF-IR.
DNA Synthesis in R− Cells.
The next aim of this investigation was to determine whether GEP uses the same signaling pathways as the IGF-IR to stimulate growth of MEF. As stated in the “Introduction,” IGF-I needs the cooperation of other growth factors to stimulate mitogenesis, but when the IGF-IR is overexpressed, even modestly (6), IGF-I induces mitogenesis in the absence of other growth factors. Therefore, one possibility is that GEP may activate both pathways that are thought to be sufficient for IGF-IR-mediated mitogenesis (see “Discussion”). For this purpose, we took advantage of the fact that PDGF and EGF, although sharing some signaling pathways with the IGF-IR (see “Discussion”), cannot stimulate the proliferation of R− cells. EGF cannot even induce DNA synthesis in R− cells (37). PDGF can induce DNA synthesis but cannot make R− cells divide (41). We therefore compared in R− cells GEP signaling with signaling by either PDGF or EGF. We first confirmed that GEP and PDGF induce DNA synthesis in R− cells, whereas EGF does not (Fig. 6). In all cases, the cells tested were parental R− cells made quiescent in SFM. They were then stimulated with either CM from R−cells or CM from clones of R− cells expressing GEP, or EGF, or PDGF. The fraction of cells synthesizing DNA was determined by BrdUrd labeling.
Signaling Pathways of the GEP: The Shc/MAPK Pathway.
For these experiments, to obtain a good activation of the putative signaling pathways, the R−/GEP cells were tested not only in SFM but also after stimulation with CM of R−/GEP cells. As a control, in all these experiments, we have used CM from R− cells, untransfected or carrying only the vector. Because this CM was totally inactive, we are showing its effect (or rather, lack of effect) in some experiments but not in others. In all instances, the CM from R− cells had the same effect as SFM, i.e., no stimulation at all. In a previous report (38), we had shown that IRS-1 is not phosphorylated by treatment of R− cells with GEP, but that MAPKs were activated. It is generally assumed that the MAPK pathway, in cells stimulated with IGF-I, begins at the level of the Shc proteins, although a component of the pathway may also originate through IRS-1 (see the “Introduction”). The first question we asked was whether Shc may be tyrosyl phosphorylated by GEP. Fig. 7 A, upper panel, shows that tyrosyl phosphorylation of Shc is already detectable in R−/GEP cells growing in SFM (Lane 4), and that a slight increase occurs if the cells are treated with CM from R−/GEP cells, clone 14 (Lane 5). In neither case, however, is the stimulation as good as that obtained with EGF (Lane 3), which also phosphorylates the p66 isoform of Shc. The middle panel gives the amounts of Shc proteins immunoprecipitated, and the lower panel shows the amounts of Grb2 that coprecipitate with phosphorylated Shc (13).
The activation of MAPK was repeated at times longer than those investigated previously (38) and compared again to the effect of EGF and PDGF (5, 37, 41). R− GEP cells and control cells expressing only the vector used for the transfection of GEP cDNA were used and stimulated with CM from R−/GEP cells. Fig. 7,B shows that MAPKs are already activated in R− epi14 cells, even in SFM (Lane 3, compare with Lane 1, in which R− cells were transfected with the empty vector). The addition of CM from the same cells further increases MAPK activation, with a maximum at 10 min (Lane 4). MAPK activation remains elevated to the sustained levels of cells in SFM. R−/neo cells are also strongly stimulated by CM from R− epi14 cells (Lane 2). The lower row of Fig. 7 B gives the protein levels; only Lane 2 is low, emphasizing the strong stimulation that the GEP-CM can exert on the activation of MAPK.
In contrast, when the experiment was repeated on R−/vector cells stimulated with EGF, MAPKs were again sharply activated at 10 min but decreased rapidly at 30 min, returned to very low levels by 60 min, and remained low (Fig. 7,C, middle diagram). For convenience, in Fig. 7,C, we give the densitometric measurements of MAPK activation, normalized to the total protein amounts in each lane. The last measurement in the middle diagram of Fig. 7,C is serum stimulation, given as a comparison. When the same cells were treated with PDGF (Fig. 7,C, right diagram), the activation of MAPK followed the same pattern of R− cells stimulated with GEP (left diagram in Fig. 7 C). It increased sharply and remained elevated for at least 4 h.
These experiments, therefore, indicate that the MAPK pathway is similarly activated in R− cells by GEP and PDGF, whereas EGF cannot give a sustained activation of MAPK in these cells. These results are compatible with the different effects of these three growth factors on DNA synthesis in R− cells (Fig. 6).
The Akt Pathway.
In preliminary experiments, we had found that the stimulation of R− cells by GEP is inhibited by inhibitors of PI3-kinase (not shown), as reported frequently for other 3T3 types of cells and different growth factors (14, 15). We therefore examined the phosphorylation of Akt by GEP. Akt is already phosphorylated in GEP-expressing cells in SFM (Fig. 8,A, Lane 2), approximately to the same extent as the activation of Akt in R−/neo cells by 10% serum (Lane 4). The addition of CM from GEP-producing R− cells has only a modest effect on the constitutive Akt activation (Lane 3). Protein amounts are shown in the lower row. In Fig. 8 B, we show that PDGF and EGF also activate the Akt pathway in R− cells, in fact, even better than GEP.
To confirm that GEP activates the PI3-kinase/Akt/p70 pathways (11), we studied the activation of p70S6K in R− cells (Fig. 9,A). p70S6K activation is clearly increased in cells expressing GEP (Fig. 9,A, Lane 6) and/or in cells stimulated with CM from GEP-expressing R− cells (Fig. 9,A, Lane 7). This is true whether we used mobility shift (not shown) or phospho-specific antibodies to p70S6K (Fig. 9,A). We also tested the activation of p70S6K by PDGF (Fig. 9,A, Lane 4) and EGF (Fig. 9 A, Lane 3). Both of these growth factors activate p70S6K. The protein amounts are given in the lower row.
Although MAPK activation seems to separate EGF signaling from GEP and PDGF signaling, the last two growth factors have thus far been indistinguishable. Yet, GEP induces cell division in R− cells, whereas PDGF does not. In a previous report (57), we studied R− cells expressing different numbers of IGF-IR (6). Cells with 15 × 103 IGF-IR/cell (R508 cells) entered DNA synthesis but did not divide when stimulated with IGF-I. Cells with 30 × 103 or more receptors/cell (for instance, R600 cells) divided and grew. In that report, we showed that signal transduction was similar in both R508 and R600 cells, except for the activation of cyclin B1, which remained elevated in R600 cells even at 48 h, while decreasing to quiescence levels in R508 cells. Fig. 10 shows the activation of cyclin B1 by GEP and PDGF in R− cells. CM from R−/GEP cells induces cyclin B1 expression, which remains elevated even at 48 h after stimulation (Fig. 10,A). PDGF also induces expression of cyclin B1 (Fig. 10 B), but at 48 h, cyclin B1 levels are the same as at 0 time. In both panels, the last lane gives the cyclin B1 levels at 48 h after stimulation with 10% serum (which is fully mitogenic in R− cells).
Our results can be summarized as follows: (a) the GEP is a potent mitogenic growth factor, that, when stably expressed in R− 3T3 cells, can induce continuous proliferation. GEP remains, thus far, the only purified growth factor that can stimulate the growth of R− cells, i.e., of cells devoid of IGF-IR; (b) despite its mitogenicity, GEP is incapable of transforming 3T3 cells, even 3T3 cells with endogenous IGF-IR; (c) GEP is also incapable of protecting R− cells from anoikis, a form of apoptosis that occurs when cells are denied attachment to a substratum; (d) in GEP-expressing cells, there is a constitutive activation of the pathways that are activated in 3T3 cells (and other cell types) when an overexpressed IGF-IR is stimulated by IGF-I. The key event for cell division is the induction of cyclin B1, whereas the activation of MAPK and the Akt pathway may be sufficient for DNA synthesis but not for cell division (57). This finding and the other findings in this report on the stimulation of these pathways by PDGF and EGF explain why GEP, but not EGF and PDGF, can stimulate the growth of cells devoid of IGF-IR.
To understand the significance of these findings, one has to keep in mind that normal cells are usually growth regulated, i.e., require growth factors for proliferation. Many dysregulated cells (usually cancer cells) can grow in the absence of growth factors, often because they produce their own growth factors. But normal cells cannot grow in SFM and usually require more than one growth factor for proliferation. 3T3 MEFs are often taken as the paradigm of growth-regulated cells (see the “Introduction”). When the IGF-IR is, even slightly, overexpressed, MEFs can fully grow in SFM supplemented solely with IGF-I (5, 8). It is therefore reasonable to ask why GEP, by itself, is mitogenic in R− cells, which have no IGF-IR. Although R− cells can be grown in 10% serum, they are refractory to stimulation by purified growth factors, singly or in combination. The growth factors that were tested include PDGF and EGF (and, of course, IGF-I), which are the most commonly used growth factors for MEFs of the 3T3 type (1, 2, 3, 30, 34, 41). R− cells, incidentally, were developed using a 3T3 protocol (33) and can therefore be considered as a 3T3-like cell line. At the time R− cells were developed from mouse embryos with a deletion of the IGF-IR genes (35, 36), a control cell line was also developed from wild-type littermates, W cells. W cells have the typical growth-regulated characteristics of 3T3 cells (for instance, BALB/c 3T3 cells). In a previous report, one of our laboratories showed that purified GEP from BRL-3A cells stimulated the growth of R− cells. Now, using a stably expressed GEP cDNA, we confirm that its production causes R− cells to grow in SFM.
Although GEP is mitogenic in R− cells, it cannot replace some other functions of the IGF-IR. Thus, it does not protect R− cells from anoikis, which is a form of apoptosis caused by denying cells attachment to a substratum (49, 51, 52, 53). In R− cells stably transfected with a wild-type IGF-IR cDNA, IGF-I is sufficient for their survival in SFM in polyHEMA plates (49, 50). By using several mutants of the IGF-IR, we have been able to show that protection from anoikis by the IGF-IR (58) correlated with its protective effect against other forms of apoptosis, such as okadaic acid-induced apoptosis (59) and apoptosis induced by growth factor withdrawal (60). It seems, therefore, that GEP, per se, cannot replace the IGF-IR requirement for survival. This is a very interesting finding, because, thus far, mitogenicity and survival by the IGF-IR have been closely correlated, again using IGF-IR mutants for both tests (see above references). GEP seems capable of imitating the mitogenic signaling of the IGF-IR but not its survival signaling. An objection may be offered that the IGF-IR is often overexpressed to demonstrate its antiapoptotic effect. Although this is true in some cases, it is not always true, because IGF-I can protect from apoptotic injuries, even cells with a physiological number of IGF-IR (18).
GEP is, in these experiments, nontransforming. It is not surprising that GEP is nontransforming in R− cells. R− cells are very refractory to transformation by a variety of cellular and viral oncogenes (18); indeed, thus far, the only oncogene that can transform R− cells is v-src (22). An activated c-src, incidentally, cannot transform R− cells. However, GEP cannot transform BALB/c 3T3 cells, which have a normal level of IGF-IR, typical of MEFs, ∼15 × 103 receptors/cell. BALB/c 3T3 cells, like most other rodent cells, can be easily transformed by viral and cellular oncogenes; indeed, they are susceptible to spontaneous transformation. This finding is seemingly in contrast to the report by Zhang and Serrero (39), who inhibited the tumorigenesis of teratoma PC cells by transfecting them with an antisense cDNA to GEP. However, in examining the role of any molecule in cell transformation, one often finds a difference between being required and being sufficient. A survey of the literature shows that there are many instances in which antisense strategies or dominant-negative mutants of various transducing molecules can inhibit transformation and even reverse it, although their activation may not be sufficient to cause transformation. We have no explanation to offer on why GEP, although mitogenic, cannot protect R− cells from anoikis and cannot transform either R− cells or BALB/c 3T3 cells.
It should be less difficult to determine why GEP, although mitogenic, cannot protect from apoptosis, and experiments are in progress on this topic. As a hypothesis, we suggest that stimulation of mitogenesis of cells in monolayer cultures by GEP may also use the integrin signaling that occurs when cells are attached to a substratum (61, 62). When testing for anoikis, the cells are in suspension and may be lacking this auxiliary signal. Still, the IGF-IR can protect R− cells from anoikis (49, 50), indicating that for protection from anoikis, the IGF-IR may yet use another signaling pathway not shared with GEP.
GEP, as expected, is a very strong promoter of wound healing, as tested even in vitro under the conditions described in this report. Its effect was really remarkable. In several studies of a number of slides, we constantly observed that R− cells expressing GEP proliferated actively in SFM, closing the wound, whereas R− cells, as expected, were totally incapable of doing so. This activity of GEP may have some interesting clinical applications.
This investigation had two aims. The first one, discussed above, was about other possible functions of GEP that could overlap with the activated IGF-IR. The second aim was to learn how GEP can stimulate R− cells, i.e., how it can substitute the mitogenic signaling of the IGF-IR. Our hypothesis was that GEP should have on R− cells the same effect of IGF-I in cells overexpressing the IGF-IR. Although the mitogenic pathways originating from the IGF-IR (or the insulin receptor) are complex (8, 17), Hugl et al. (7) have proposed a simplified scheme in which two major pathways are sufficient to stimulate insulin- or IGF-I-mediated mitogenesis (and survival). One pathway proceeds from IRS-1 to PI-3 kinase, Akt, and p70S6K (14, 15), whereas the other derives from Shc proteins, through Grb2, Sos, Ras, Raf, and MAPK. The latter pathway could also be activated by the binding of Grb2 to IRS-1 (17). When both pathways are activated by IGF-I, the cells enter DNA synthesis and divide. Although this is a simplified scheme, it is sufficient, at least as a starting point to determine whether GEP uses similar mitogenic pathways. In a previous report, we had shown that GEP did not tyrosyl phosphorylate IRS-1, although it induced MAPK activation (38). However, there are by now several reports indicating that, to have an effect, MAPK activation should be sustained (63). Only when MAPK activation is sustained, is there translocation to the nucleus and activation of the genetic program that leads to cell cycle progression (64). GEP seems to activate the same pathways that are activated by the IGF-IR. It induces tyrosyl phosphorylation of Shc and sustained activation of MAPK, up to 4 h. Despite the fact that GEP does not activate IRS-1, it requires a functional PI3-kinase (not shown), and it activates Akt and p70S6K. Thus, if we accept the simplified scheme of Hugl et al. (7), GEP can be mitogenic in R− cells because it induces, by itself, the same signaling pathways that are induced by an overexpressed IGF-IR activated by its ligands. Both MAPK and p70S6K are fully activated by GEP in the absence of other growth factors.
We have compared the signal-transducing pathways in R− cells of GEP to those induced by PDGF and EGF. The results may explain why these two growth factors cannot stimulate R− cells, although GEP can. PDGF is known to activate the PI3-kinase/p70 pathway (65, 66). PDGF can activate both the MAPK and p70S6K pathways in R− cells, which explains its ability to induce DNA synthesis in these cells. But GEP in R− cells and IGF-I in cells overexpressing the IGF-IR must provide additional information that allows the cells to go through G2 and divide. Indeed, it has been established (34) that continuous stimulation by IGF-I is required for the progression of cells from G2 through mitosis, an observation that has been confirmed independently (67). A clue was offered by the report of Reiss et al. (57) on the induction of cyclin B1. Cyclin B1 levels determine the entry and the exit of cells from mitosis (68). If the cells stop in G2, cyclin B1 levels (and cdc2 kinase activity) decrease to prestimulation levels, whereas they remain elevated if the cells progress from G2 into mitosis. Our experiments confirm these findings: cyclin B1 levels decrease at 36 h in R− cells stimulated with PDGF but not in cells stimulated with GEP. GEP, therefore, like IGF-I in cells overexpressing the IGF-IR, activates the full cell cycle program, whereas PDGF (at least in R− cells) offers only a partial activation (DNA synthesis).
In conclusion, we show that a plasmid expressing GEP, stably transfected into R− cells, causes them to grow in SFM, confirming the results of Xu et al. (38) with the purified protein. GEP also promotes vigorous wound healing. GEP, however, cannot transform MEFs, under the conditions used, and cannot protect them from anoikis.
The mitogenicity of GEP in R− cells can be explained by the fact that it activates the two pathways that are sufficient for the mitogenicity of an overexpressed IGF-IR. When the IGF-IR is not overexpressed, as in BALB/c 3T3 cells or human diploid fibroblasts, a combination of PDGF and IGF-I (1) or EGF and IGF-I (3) is required for cell proliferation. Significantly, PDGF increases IGF-IR number (6), whereas EGF stimulates IGF-I production (5). We can therefore say that GEP can replace both an overexpressed IGF-IR or a combination of the growth factors that makes 3T3 cells grow. In other words, GEP can, by itself, activate the cell cycle machinery of MEFs, under conditions that are forbidding to other single growth factors. It does so, apparently, by activating the two pathways that are activated by an overexpressed IGF-IR.
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.
Supported by Grants CA 53484 and CA 56309 from the NIH and Grant MT-11288 from the Medical Research Council of Canada.
The abbreviations used are: SFM, serum-free medium; MEF, mouse embryo fibroblast; IGF, insulin-like growth factor; IGF-IR, IGF-I receptor; EGF, epidermal growth factor; PDGF, platelet-derived growth factor; PI3-kinase, phosphatidylinositol 3-kinase; Shc, Src homology and collagen; Grb2, growth factor receptor binding protein 2; MAPK, mitogen-activated protein kinase; GEP, granulin/epithelin precursor; MLV, murine leukemia virus; FBS, fetal bovine serum; mAb, monoclonal antibody; CM, conditioned medium; BrdUrd, 5-bromo-2′-deoxyuridine; polyHEMA, poly (2-hydroxyethyl methacrylate).