Regulation of Id2 Gene Expression by the Insulin-like Growth Factor I Receptor Requires Signaling by Phosphatidylinositol 3-Kinase*

The Id proteins play an important role in proliferation, differentiation, and tumor development. We report here that Id gene expression can be regulated by the insulin-like growth factor I receptor (IGF-IR), a receptor that also participates in the regulation of cellular proliferation and differentiation. Specifically, we found that the IGF-IR activated by its ligand was a strong inducer of Id2 gene expression in 32D murine hemopoietic cells. This activation was not simply the result of cellular proliferation, as Id2 gene expression was higher in 32D cells stimulated by IGF-I than in cells exponentially growing in interleukin-3. The up-regulation of Id2 gene expression was largely dependent on the presence of insulin receptor substrate-1, a major substrate of the IGF-IR and a potent activator of the phosphatidylinositol 3-kinase (PI3K) pathway. The role of PI3K activity in the up-regulation of Id2 gene expression by the IGF-IR was confirmed by different methods and in different cell types. In 32D cells, the up-regulation of Id2 gene expression by the PI3K pathway correlated with interleukin-3 independence and inhibition of differentiation.

The Id proteins play an important role in proliferation, differentiation, and tumor development. We report here that Id gene expression can be regulated by the insulin-like growth factor I receptor (IGF-IR), a receptor that also participates in the regulation of cellular proliferation and differentiation. Specifically, we found that the IGF-IR activated by its ligand was a strong inducer of Id2 gene expression in 32D murine hemopoietic cells. This activation was not simply the result of cellular proliferation, as Id2 gene expression was higher in 32D cells stimulated by IGF-I than in cells exponentially growing in interleukin-3. The up-regulation of Id2 gene expression was largely dependent on the presence of insulin receptor substrate-1, a major substrate of the IGF-IR and a potent activator of the phosphatidylinositol 3-kinase (PI3K) pathway. The role of PI3K activity in the up-regulation of Id2 gene expression by the IGF-IR was confirmed by different methods and in different cell types. In 32D cells, the up-regulation of Id2 gene expression by the PI3K pathway correlated with interleukin-3 independence and inhibition of differentiation.
The Id proteins are helix-loop-helix proteins that form heterodimers with a large family of other helix-loop-helix proteins, mostly transcriptional activators, that have a basic region in addition to the helix-loop-helix region (1). These heterodimers cannot bind to DNA because the Id proteins lack a DNAbinding region; Id proteins therefore negatively regulate the DNA-binding capacity of other basic helix-loop-helix proteins (1). There are at least four Id proteins encoded by individual genes, which, in humans, are located on different chromosomes (2). Id gene expression is elevated in undifferentiated cycling cells and tumor cell lines (3,4), and high levels of Id gene expression inhibit the differentiation of a variety of cell types (3,(5)(6)(7). Id gene expression is also cell cycle-regulated and has been implicated in playing a major role in the G 1 -to-S transition (2,8,9). The Id1 protein has been reported to inhibit mammary cell differentiation (6), to promote mammary epithelial cell invasion (10), and to increase the aggressive phenotype of human breast cancer cells (11). A connection between N-Myc, the retinoblastoma protein, and Id2 has been recently established in neuroblastoma cells (9). Id proteins are also required for angiogenesis and vascularization of tumor xenografts (12).
Id gene expression is up-regulated by serum (8,9), by platelet-derived growth factor (13), and by some cytokines that induce cellular differentiation (14,15). The dual role of Id proteins in proliferation and differentiation has prompted us to examine their regulation by the insulin-like growth factor I receptor (IGF-IR), 1 which is also involved in the proliferation and differentiation of cells. We have chosen for these studies the 32D murine myeloid cell line (16), in which induction of differentiation requires a short but intense period of cell proliferation, as it happens for other hemopoietic cells induced to differentiate (16 -18). This dual response has been interpreted as indicating that differentiating growth factors send two simultaneous signals, one for proliferation and one for differentiation, with the latter eventually prevailing.
The IGF-IR offers an ideal paradigm for studying how growth factors regulate the balance between cell proliferation and cell differentiation. In many cell types (mouse embryo fibroblasts like 3T3 cells, human diploid fibroblasts, some epithelial cells, etc.), the IGF-IR sends an unambiguous mitogenic signal (19). However, in other cell types, IGF-I and IGF-II can stimulate either proliferation or differentiation, or both (20). These contradictory signals of the IGF-IR have been studied to advantage in 32D cells, a cell line that is interleukin-3 (IL-3)dependent for growth (16). In the absence of IL-3, 32D cells undergo apoptosis (21). The expression of a human IGF-IR in 32D cells (32D IGF-IR cells) prevents apoptosis caused by IL-3 withdrawal (22,23). In fact, 32D IGF-IR cells are stimulated to grow exponentially by IGF-I for ϳ48 h (22)(23)(24). Then, the cells stop growing and begin to differentiate along the granulocytic pathway (25). Thus, 32D IGF-IR cells recapitulate the program induced in hemopoietic cells by other differentiating growth factors (see above).
A characteristic of 32D cells is that they do not express insulin receptor substrate-1 (IRS-1) and IRS-2 (25,26). The IRS proteins are major substrates of both the IGF-IR and the insulin receptor and play an important role in the mitogenic signaling of both receptors (27). Ectopic expression of IRS-1 in 32D IGF-IR cells (32D IGF-IR/IRS-1 cells) causes inhibition of differentiation (25). In fact, 32D IGF-IR/IRS-1 cells become permanently IL-3-independent and form tumors in mice (24).
We have taken advantage of this dual signaling from the IGF-IR (differentiation in the absence of IRS-1 and continuous proliferation in its presence) to investigate the signal transduction pathways leading to Id gene expression, specifically Id1 and Id2. We found that the IGF-IR can induce the expression of Id2 mRNA and proteins, especially when IRS-1 is present. Since IRS-1 is a strong activator of PI3K (27,28), we investigated the role of PI3K in the expression of Id2 mRNA. Using inhibitors of PI3K and a mutant IRS-1 that fails to activate PI3K, we show that, in 32D IGF-IR cells, PI3K activation plays an important role in the up-regulation of Id2 gene expression.
We have confirmed these findings in two different cell types, 293T and LNCaP cells. With 293T cells (29), we transiently transfected a constitutively active p110␣ subunit of PI3K (30), which caused an increase in Id2 expression. LNCaP cells are human prostatic cancer cells (31) with a frameshift mutation of PTEN, an inhibitor of the PI3K pathway (see below). Re-expression of PTEN inhibits IGF-I-mediated up-regulation of Id2 proteins. Our results clearly show that the IGF-IR activated by its ligand is capable of inducing Id2 gene expression through the IRS-1/PI3K signaling pathway. Although other pathways probably also lead to Id2 gene expression, this is, to our knowledge, the first signal transduction pathway to be identified. Significantly, these findings also offer an explanation for the dual signaling of the IGF-IR in these cells.

EXPERIMENTAL PROCEDURES
Plasmids and Retroviral Transduction-A truncated mouse IRS-1 was generated that encodes the first 309 amino acid residues and comprises the pleckstrin homology (PH) domain and the phosphotyrosine-binding (PTB) domain. For its construction, wild-type IRS-1 was used as a template for PCR amplification. The sequence of the sense primer was 5Ј-ATACCGTTACACaagcttGGCGCAGTTACCT-CGTCCTTCGG-3Ј (positions 301-327 of the mouse IRS-1 genome (GenBank TM /EBI accession number X69722); the sequence of the Hind-III restriction site is indicated in lowercase letters, and the overhang sequence of the primer is underlined). The sequence of the antisense primer was 5Ј-TTCATAGCATTTCGTCATTATaagcttTTTCCCACCCA-CCATTCAGGCAGG-3Ј (positions 1418 to 1394 of the mouse IRS-1 genome (accession number X69722); the sequence of the HindIII restriction site is indicated in lowercase letters, and the overhang sequence of the primer is underlined). The PCR conditions were as follows: a denaturation step of the template at 94°C for 1 min, followed by a 1-min interval at 52°C to allow for the annealing of the primers to the template and a 3-min incubation at 72°C for polymerase elongation of the primers. This cycle was repeated 35 times. At the end of these 35 cycles, an additional incubation at 72°C for 15 min was included to allow completion of the amplification. The PCR product was electrophoresed on a 1% agarose gel. The correct size product was excised from the gel and purified with a gel extraction kit (QIAGEN Inc.) following the manufacturer's recommendations. The end PCR product was in 50 l of water and was digested with the HindIII restriction enzyme. The HindIII-digested PCR product was gel-purified again and ligated into the HindIII cloning site of a murine stem cell virus based retroviral vector carrying the puromycin resistance gene. The HindIII PCR fragment was sequenced to monitor the mutation. For details of the retroviral vector, see Romano et al. (32).
The PI3K constructs were hemagglutinin-tagged and included wildtype p110␣ (pLHA110) and constitutively active p110␣ CAAX (pLHACAAX) and were a kind gift of Christian Sell. The p110␣ sequences are under the control of the cytomegalovirus promoter, and the constructs are described by Didichenko et al. (30).
Cell Lines-32D IGF-IR cells were derived from the 32D murine hematopoietic cell line clone 3 (16), stably transfected with a plasmid expressing the human IGF-IR cDNA (25). 32D IGF-IR/IRS-1 cells are 32D IGF-IR cells expressing wild-type IRS-1 (25), and 32D IGF-IR/PH/ PTB cells are 32D IGF-IR cells expressing a truncated form of IRS-1 under the control of the cytomegalovirus promoter. 293T cells were originally obtained from the laboratory of David Baltimore (29). LNCaP, LNCaP/PTEN, and LNCaP/PTEN/IRS-1 cells have been described by Reiss et al. (31). LNCaP/PTEN/IGF-IR cells were generated by transducing LNCaP/PTEN cells with a retrovirus expressing the human IGF-IR cDNA. Selection was carried out with 2 g/ml puromycin and 500 g/ml G418 (both from Life Technologies, Inc.). All these cell lines are mixed populations.
32D cells were grown in RPMI 1640 medium supplemented with 10% heat-inactivated fetal bovine serum (Life Technologies, Inc.), 10% WEHI cell-conditioned medium (as a source of IL-3), 2 mM L-glutamine (Life Technologies, Inc.), and the required antibiotic to maintain the selective pressure (250 g/ml G418 for 32D IGF-IR cells, 250 g/ml G418 plus 250 g/ml hygromycin (Calbiochem) for 32D IGF-IR/IRS-1 cells, or 250 g/ml of G418 plus 1 g/ml puromycin for 32D IGF-IR/PH/ PTB cells). For brevity, the WEHI cell-conditioned medium will be referred to as IL-3. In the experiments for Id1 or Id2 detection, exponentially growing cells (in complete medium with 10% fetal bovine serum and IL-3) were washed in Hank's balanced solution to remove IL-3 and incubated for the indicated times in medium with 10% fetal bovine serum plus 50 g/ml IGF-I.
LNCaP cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum. In the experiment for Id2 detection, cells were serum-starved in RPMI 1640 medium and 0.1% bovine serum albumin (Sigma) for 48 h and then 50 ng/ml IGF-I was added for the indicated times.
293T cells were grown in Dulbecco's modified Eagle's medium and 10% fetal bovine serum. For the transient expression experiments, 293T cells (29) were transfected with the PI3K constructs (see above) using the FuGENE transfection reagent (Roche Molecular Biochemicals) according to the manufacturer's instructions. After 48 h, the cells were harvested, and lysates were prepared as previously described (25).
Northern Blots-For the detection of the Id1 and Id2 mRNAs, exponentially growing cells were washed three times and seeded in IL-3-free medium (RPMI 1640 medium containing only 10% heat-inactivated fetal bovine serum and L-glutamine), and then IGF-I (50 ng/ml) was added. Cells incubated with IL-3 served as controls. In some experiments, an inhibitor was added (30 M LY294002; BIOMOL Research Labs Inc.) for 15 min prior to the stimulation with IGF-I. At the indicated time points, cells were collected, and total RNA was extracted using the RNeasy kit (QIAGEN Inc.) following the manufacturer's instructions. Northern blotting was carried out by standard techniques. The labeled probes used were a 250-base pair PCR product for Id1 (spanning bases 259 -530) and the full-length cDNA for Id2.
PI3K Activity Assay-Exponentially growing cells were collected, washed three times, and starved for 5 h in serum-free medium (RPMI 1640 medium with 0.1% bovine serum albumin and L-glutamine). After starvation, cells were stimulated with 50 ng/ml IGF-I for 5 min and lysed. One milligram of proteins was then immunoprecipitated with an anti-Tyr(P) monoclonal antibody (PY20, Transduction Laboratories) overnight at 4°C, and the associated PI3K activity was assessed by incubating the immunoprecipitate with phosphatidylinositol and [␥-32 P]ATP for 10 min at 37°C. The products of the kinase reaction were then visualized by autoradiography and quantified by the radioactivity of the excised bands in a scintillation counter.
Growth and Differentiation Analyses-These analyses were carried out by routine methods, as already described in detail for 32D cells in previous works from this laboratory (23)(24)(25). Cell growth was also assessed as the fraction of cells in S phase by 5-bromo-2Ј-deoxyuridine (BrdUrd) incorporation using the BrdUrd labeling kit (Roche Molecular Biochemicals). Exponentially growing cells were collected, washed three times, and resuspended in IL-3-free medium or supplemented with 50 ng/ml IGF-I or 10% IL-3-conditioned medium. Cells were seeded at a density of 5 ϫ 10 4 cells/ml on 6-well multiwell plates. At the indicated time points, 10 M BrdUrd was added to the medium for 1 h before collecting the cells. Cytospins were performed, and slides were fixed for 20 min in 3% paraformaldehyde, permeabilized for 5 min in 0.2% Triton, treated for 5 min with 1.5 N HCl, and subsequently incubated for 30 min at 37°C with the primary anti-BrdUrd monoclonal antibody and for 30 min at room temperature with the secondary Texas Red-conjugated anti-mouse antibody. Hoechst 33258 (Sigma) staining of the nucleus was performed for 5 min at room temperature before applying Vectashield mounting medium (Vector Laboratories, Inc., Burlingame, CA). The BrdUrd incorporation index was determined using a Zeiss microscope working in epifluorescence mode (magnification ϫ 500). In randomly selected fields, at least 300 cells/slide were counted.

RESULTS
An intriguing aspect of these experiments is that in the first 48 h after IL-3 withdrawal followed by IGF-I stimulation, 32D IGF-IR cells (eventually undergoing differentiation) and 32D IGF-IR/IRS-1 cells (transformed) could not be distinguished from each other in terms of cell proliferation. Both cell lines double in number in each 24-h period (22-24) (see below), although their fates diverge in opposite directions soon after. Furthermore, all 32D-derived cell lines, even those expressing a disabled IGF-IR grow exponentially in IL-3 (22,23,25). Therefore, any change in Id gene expression in the first 24 h after shifting from IL-3 to IGF-I would not be simply a reflection of the proliferative status of the cells. We have investigated Id mRNA expression in 32D IGF-IR cells and 32D IGF-IR/IRS-1 cells during this period, limiting ourselves to Id1 and Id2 gene expression because Id3 and Id4 are not expressed in parental 32D cells (33).
IGF-I Stimulates an Increase in Id2 mRNA Levels- Fig. 1A shows Id2 mRNA levels in Northern blots from 32D IGF-IR cells and 32D IGF-IR/IRS-1 cells. Id2 RNA was barely detectable in 32D IGF-IR/IRS-1 cells growing in IL-3, and its levels increased sharply when the cells were shifted from IL-3 (Time 0 lane) to IGF-I, with a peak at 6 h. Significantly, the Id2 RNA levels were lower in IL-3. Even at 24 h, Id2 RNA levels were higher in cells incubated in IGF-I than in cells incubated in IL-3, although the cells were exponentially growing in both conditions. The levels of Id2 RNA were much lower in 32D IGF-IR cells, already indicating that IRS-1 plays a crucial role in determining Id2 RNA levels after IGF-I stimulation. However, a very modest but reproducible increase in Id2 RNA could also be observed in 32D IGF-IR cells (see also below). These experiments were repeated several times by different operators, with similar results.
The results with Id1 mRNA (Fig. 1A, middle panel) were different. In fact, it was difficult to see a clear difference in Id1 RNA levels between 32D IGF-IR cells and 32D IGF-IR/IRS-1 cells. There was perhaps a peak in Id1 RNA levels in the latter cells at 6 h after IGF-I stimulation, but the effect was modest. In addition, Id1 RNA levels were already easily detectable in cells in IL-3, and shifting to IGF-I did not seem to increase them. The results indicate that regulation of Id1 and Id2 RNAs differs in 32D-derived cells. In the following experiments, we focused mostly on Id2 gene expression.
To demonstrate that the IGF-IR can indeed up-regulate Id2 gene expression, we compared Id2 RNA levels in 32 IGF-IR/ IRS-1 cells in 10% serum and in 10% serum supplemented either with IL-3 or IGF-I for 6 h (Fig. 1B). Only the addition of IGF-I caused a convincing increase in Id2 RNA levels (third lane). The experiments in Fig. 1 (A and B) are a clear indication that Id2 gene expression is specifically up-regulated by IGF-I. In Fig. 1C, we show that Id2 protein levels also increased in 32D IGF-IR/IRS-1 cells. Id2 protein (ϳ14 kDa) amounts increased by 6 h after shifting the cells from IL-3 to IGF-I, and although they eventually decreased, they were still high at 72 h. Id2 protein levels were much less elevated in 32D IGF-IR cells, although they were slightly higher in IGF-I than in IL-3. These experiments conclusively demonstrate the ability of the IGF-IR, activated by its ligand, to up-regulate the expression of Id2 RNA and protein, especially in the presence of its substrate IRS-I. In subsequent experiments with 32D cells, we focused on RNA levels to emphasize the fact that changes in Id2 gene expression occur at a time when all cell lines examined are exponentially growing.
PI3K Is Required for Strong Induction of Id2 mRNA-Wild-type IRS-1 is known to send a powerful mitogenic stimulus through the PI3K, Akt/protein kinase B, and p70 S6K pathway (28). Since parental 32D and 32D IGF-IR cells do not express IRS-1 (25,26), the results of Fig. 1 suggest that PI3K activity may be crucial for the up-regulation of Id2 RNA in 32D-derived cells. As a first approach, we determined the levels of PI3K activity in the two cell lines. As expected, PI3K activity, after IGF-I stimulation, was substantially higher in 32D IGF-IR/IRS-1 cells than in 32D IGF-IR cells ( Fig. 2A). There may be a slight increase in the latter cell line after IGF-I stimulation. This slight increase is compatible with our previous observation that IGF-I causes a modest but reproducible increase in Akt activation even in 32D IGF-IR cells (25). We then compared Id2 PTB domains and lacks the binding domains for PI3K.This truncated IRS-1 is only 309 amino acids long and can be detected only with an antibody to the amino terminus of IRS-1 (see "Experimental Procedures"). It has been reported that the PH/PTB mutant does not activate PI3K (26). The effects of this mutant IRS-1 on Id expression are shown in Fig. 3A. For Id2 RNA, we compared 32D IGF-R/IRS-1 cells with 32D IGF-IR/ PH/PTB cells. The mutant IRS-1 failed to elicit a strong induction of Id2 RNA (Fig. 3A, upper panel). Very little differences were noted when Id1 RNA levels were examined in the same cells, again indicating that Id1 RNA levels, in this model, are regulated by different pathways than Id2 RNA levels. Fig. 3B shows that the PH/PTB mutant was well expressed in two separate mixed populations derived from 32D IGF-IR cells. In 32D IGF-IR/PH/PTB cells, both PI3K activity and Akt/protein kinase B phosphorylation were the same as in 32D IGF-IR cells (data not shown). The results therefore confirm that PI3K plays a major role in IGF-I-mediated up-regulation of Id2 gene expression.

Growth and Differentiation of 32D IGF-IR/PH/PTB Cells-
The effect of IGF-I on the growth and differentiation response of 32D IGF-IR/PH/PTB cells has never been reported. Yenush et al. (26) have noticed a modest but reproducible proliferative stimulus originating from the PH/PTB domain of IRS-1 in 32D cells overexpressing the insulin receptor. We investigated the ability of the PH/PTB mutant of IRS-1 to send proliferative and/or differentiation signals in the context of the 32D cells expressing the IGF-IR. Fig. 4 (A and B) compares the growth and differentiation of the three cell lines: 32D IGF-IR, 32D IGF-IR/IRS-1, and 32D IGF-IR/PH/PTB. As already reported, 32D IGF-IR cells (black bars) grew for a period of time (A), but then began to differentiate (B), and the number of cells no longer increased, whereas 32D IGF-IR/IRS-1 cells (hatched bars) continued to grow. The percentage of differentiating cells reached 40% by day 6 in 32D IGF-IR cells, whereas it remained at background levels in 32D IGF-IR/IRS-1 cells, confirming the results of Valentinis et al. (25). 32D IGF-IR/PH/PTB cells (stippled bars) fell somewhere in between in terms of cell proliferation, which was reproducibly better than that of 32D IGF-IR cells, but not as brisk as that of 32D IGF-IR/IRS-1 cells. The 32D IGF-IR/PH/PTB cells differentiated as well as the parental 32D IGF-IR cells. It seems therefore that 32D IGF-IR/PH/PTB cells have an increased stimulus for cell proliferation, but that the differentiation program is not extinguished. Therefore, in this model, PI3K modulates not only the expression of the Id2 genes, but also the differentiation program. In the absence or with a marked reduction of PI3K activity, 32D IGF-IR/PH/PTB cells differentiate, even in the presence of a mild proliferative stimulus.
We have looked at the incorporation of BrdUrd into these three different cell lines. Table I shows the results of a 1-h pulse labeling with BrdUrd at various times after IL-3 withdrawal and IGF-I supplementation. 32D IGF-IR/PH/PTB cells incorporated BrdUrd at a percentage that was substantially higher than 32D IGF-IR cells, although not as high as the percentage obtained with 32D IGF-IR/IRS-1 cells. For completeness, it should be added that all three cell lines, as usual, grow equally well in IL-3, and all undergo apoptosis if the IL-3-depleted medium is not supplemented with IGF-I (22,24,25).
Effect of PI3K and PTEN on Id2 Gene Expression-To confirm the role of the IGF-IR and the importance of the PI3K pathway in the regulation of Id2 expression, we investigated the regulation of Id2 gene expression in two other cell lines and by two different methods. In one series of experiments, we studied Id2 gene expression in 293T cells (29) transiently transfected with wild-type p110␣ or a constitutively active mutant of the p110␣ subunit of PI3K (30). Levels of Id2 proteins were determined 48 h after transfection, and the results are shown in Fig. 5A. Both the wild-type and constitutively active p110␣ subunits caused an increase in Id2 protein levels (second and third lanes). Fig. 5B shows the levels of expression of p110␣ (with a hemagglutinin tag) in the transfected cells. Equal loading was monitored by anti-Grb2 hybridization (Fig.  5C). The experiments were also repeated with two other constructs, wild-type p110␣ and constitutively active p110␣ tagged with a Myc sequence. The results were essentially the same, except that the wild-type construct, in this case, did not produce an increase in Id2 protein (data not shown). For a third cell type, we chose LNCaP cells, a human prostatic cancer cell line with a frameshift mutation of PTEN, a tumor suppressor gene (also called MMAC1 or TEP1) identified on human chromosome 10q23 (34,35). PTEN is a phosphatase (36,37) that regulates the activity of PI3K (38 -41) and blocks cells in the G 1 phase of the cell cycle (42,43). Accordingly, Akt is constitutively activated in LNCaP cells (31,39,40,44,45) because of the PTEN mutation. Fig. 5D shows the levels of expression of the Id2 protein in four different cell lines, three of which have been previously described (31): parental LNCaP, LNCaP/PTEN, and LNCaP/ PTEN/IRS-1 cells. A fourth cell line was generated by introducing a retrovirus expressing the IGF-IR into LNCaP/PTEN cells (see "Experimental Procedures"). All these cell lines are mixed populations and were tested for IGF-I-mediated induction of Id2 gene expression. The cells were made quiescent by serum deprivation for 48 h and then stimulated with IGF-I (50 ng/ml). Fig. 5D is a Western blot for Id2 protein at various times after IGF-I stimulation. The Id2 protein was up-regulated in parental LNCaP cells by IGF-I. The presence of PTEN abrogated the response. In LNCaP/PTEN/IRS-1 cells, a slight increase (over LNCaP/PTEN cells) was detectable, as if IRS-1 were restoring the PI3K modulation of Id2. This increase, albeit extremely modest, was reproducible. We have no explanation at the present moment for the constitutive expression of Id2 in these latter cells, even before IGF-I stimulation. Fig. 5E shows the expression of PTEN in each cell line. PTEN is expressed strongly in all mixed populations, except the parental LNCaP cells, where PTEN is not detectable (31,34,35). Equal loading was monitored by standard methods (data not shown).

DISCUSSION
The novel findings of this study are summarized and discussed below. 1) Id2 gene expression is up-regulated by an activated IGF-IR both in hemopoietic cells and in other cell types. 2) In 32D cells, induction of Id2 gene expression is largely dependent on the presence of IRS-1, a major substrate of the IGF-IR. 3) Activation of the PI3K pathway (which is strongly promoted by IRS-1) is apparently the main pathway for the up-regulation of Id2 gene expression. 4) Transient expression of the wild-type or constitutively active PI3K subunit induces an increase in Id2 protein expression. 5) In LNCaP prostate cancer cells, the important role of the PI3K pathway was confirmed by the inhibition of Id2 gene expression by PTEN, an inhibitor of PI3K. 6) Up-regulation of Id2 gene expression by IGF-I is not simply due to the proliferative status of the cells, as Id2 RNA and protein levels increase when 32D cells are shifted from IL-3 to IGF-I. 7) The modulation of Id2 gene expression by the IRS-1/PI3K pathway correlates with the inhibition of IGF-I-mediated differentiation of 32D IGF-IR cells. A secondary finding is the failure of IGF-I to induce significant changes in Id1 expression.
Although there is a substantial literature on the role of Id proteins in proliferation and differentiation (see the Introduction), information on the signaling pathways regulating Id gene expression is limited. Serum, nerve growth factor, plateletderived growth factor, and some cytokines have been mentioned as regulators of Id gene expression, but no information on the signal transducing pathways leading to Id gene expression has been forthcoming. We show in this study that activation of the IGF-IR is a signal for up-regulation of Id2 gene expression. Id2 RNA levels actually increase when 32D IGF-IR/IRS-1 cells are shifted from IL-3-to IGF-I-supplemented medium. In a direct comparison, IGF-I induces Id2 RNA, whereas IL-3 fails to do so. At least in these cells, IGF-I is a more potent activator of Id2 gene expression than IL-3. This is significant, as during the period of observation, these cell lines (whether or not expressing IRS-1) are growing exponentially both in IL-3 and in IGF-I. It follows that the induction of Id2 gene expression by the IGF-IR is not simply due to the stimulation of cell proliferation and indicates a strong relationship between the IGF axis and Id2 gene expression. The increase in RNA levels is mirrored by an increase in the levels of Id2 proteins. Interestingly, Id2 protein levels (and Id2 RNA levels) are also slightly increased in 32D IGF-IR cells after shifting from IL-3 to IGF-I, suggesting that the IGF-IR can activate Id2 gene expression even in the absence of IRS-1. This effect is, however, in no way comparable to the effect obtained by the activated IGF-IR in the presence of IRS-1.  Since the presence of IRS-1 is a determinant factor in the up-regulation of Id2 RNA by IGF-I, we have naturally focused our attention on the PI3K pathway. IRS-1 is a potent activator of the PI3K pathway (27,28) and has binding sites for the p85 subunit of PI3K (26,46). The literature on IRS-1 and PI3K is abundant both for the IGF-IR and the insulin receptors (27,47), and we have confirmed the importance of IRS-1 in PI3K activation in this study. More important, we have shown clearly that PI3K is required in 32D cells for strong up-regulation of Id2 gene expression. Using inhibitors of PI3K and a mutant IRS-1 defective in the activation of PI3K, we have shown that up-regulation of Id2 gene expression is markedly inhibited in the absence of a PI3K signal (or a diminished signal, as in 32D IGF-IR cells). Interestingly, in 32D cells, PI3K is crucial not only for up-regulation of Id2 gene expression, but also for inhibition of IGF-I-mediated differentiation of these cells (24). The PH/PTB mutant of IRS-1 has a modest proliferative effect in parental 32D IGF-IR cells, confirming similar results obtained by Yenush et al. (26) with the insulin receptor. However, the PH/PTB mutant is totally incapable of extinguishing the differentiation program initiated by the IGF-IR.
The attractive aspect of the 32D model is that, as already emphasized, all three cell lines under consideration grow exponentially in the first 48 h after shifting from IL-3 to IGF-I. In IL-3, of course, all 32D-derived cells grow indefinitely. Thus, the up-regulation of Id2 RNA (already evident 3-6 h after shifting the cells to IGF-I) is independent of the actual proliferative status of the cells, but it is indicative of their fate. Our results establish a connection between Id gene expression and the PI3K pathway. To confirm the importance of the PI3K pathway in the up-regulation of Id2 gene expression, we have studied it by different methods in other cell types. We have used a constitutively active p110␣ subunit of PI3K transiently expressed in 293T cells, obtaining a modest but reproducible increase in Id2 proteins. The modest increase also obtained with wild-type p110␣ is probably due to the fact that these experiments had to be done in 10% serum, which contains, among many growth factors, also the IGFs. LNCaP cells are a prostatic cancer cell line that originated from a metastatic tumor. They express the IGF-IR, albeit at a low level, ϳ7-8 ϫ 10 3 receptors/cell (31). Peptide analogs to IGF-I inhibit their growth (48), and they respond to IGF-I with an increase in cell proliferation (49). Their requirement for IGF-I is more clearly evident under conditions of anchorage-independent growth (49). As mentioned, they have a frameshift mutation of PTEN (34,35). Because of the PTEN mutation, the PI3K pathway is constitutively activated (31,39,40,44,45). By using these cells, we have confirmed in another cell line that IGF-I upregulates Id2 gene expression and that this regulation is inhibited by PTEN, an antagonist of PI3K. Id2 protein levels are increased in LNCaP cells by IGF-I, but the increase is abrogated in cell lines (mixed populations), expressing wild-type PTEN.
The sequence of events between PI3K activation and upregulation of Id gene expression is a very complex question and remains to be fully elucidated. However, the first events after PI3K activation are well established, and they include the activation of Akt/protein kinase B and p70 S6K /p85 and the activation of a genetic program for a number of genes (50). This pathway has also been repeatedly documented in 32D-derived cells (22)(23)(24). Activation of the PI3K pathway also results in increased expression of c-Myc (51,52), and c-Myc expression is induced by IGF-I (53). In fact, the levels of c-Myc induction by IGF-I are proportional to the number of IGF-IRs/cell (54). It is therefore of great interest that Lasorella et al. (9) have recently shown that Id2 is a direct target of N-Myc in neuroblastoma cells. A connection between N-Myc and the IGF-IR in neuroblastoma cells has also been reported by Chambery et al. (55).
The two Id RNAs are somewhat different in their behavior, but this is not surprising. Differences in expression of Id1 and Id 2 RNAs during development and differentiation have already been reported (33,56,57). Our experiments confirm that there are differences in the regulation of Id1 and Id2 gene expression. Id2 RNA is clearly regulated by the IGF-IR through the PI3K pathway, whereas the results with Id1 are inconclusive. As mentioned in the Introduction, Id3 and Id4 are not expressed in 32D cells (33).
The implications of these findings are worth considering. The Id proteins have recently been proposed to play a role in malignant transformation and tumor aggressiveness (6, 9 -11). In addition, Id2 has been reported to inhibit differentiation and to enhance cellular proliferation by associating with and inactivating pRb (9,58). The activation of the PI3K pathway has also been often implicated in transformation (59,60), indeed directly in this model system, where the expression of IRS-1 makes the difference between terminal differentiation and transformation (tumor formation in mice) (24). Considering that parental 32D IGF-IR cells cannot form tumors in mice, although they grow very well in IL-3, one is tempted to speculate that Id2 expression may be more important for transformation than for mitogenesis.
There is therefore a reasonable explanation for our results, suggesting an attractive hypothesis. The IGF-IR, through IRS-1, activates the PI3K pathway (27), which results in Myc induction (51)(52)(53)(54), which induces Id2 gene expression (9). Preliminary experiments from our laboratory have indicated that up-regulation of Id2 gene expression in 32D-derived cell lines is accompanied by the induction of c-Myc RNA. 2 In LNCaP cells, the PI3K pathway is already constitutively activated by the PTEN mutation, and IRS-1 is dispensable.
In conclusion, we have shown that IGF-IR signaling can up-regulate Id2 gene expression. The up-regulation is, however, markedly increased by the presence of one of its major substrates (IRS-1) and, more specifically, by the activation of the PI3K pathway. The up-regulation of Id2 gene expression correlates with the extinction of the IGF-I-mediated differentiation program of 32D IGF-IR cells and their malignant transformation. Clearly, Id gene expression is activated by other growth factors (see the Introduction), and we expect that other signaling pathways will be found leading to up-regulation (or down-regulation) of Id gene expression. For the moment, we can say that the IGF-IR is certainly one of the receptors that can modulate Id2 gene expression and that PI3K is at least one of the pathways leading to it. These results may further open the way to an investigation of the growth factors and the signaling pathways that regulate the expression of Id protein in proliferation, differentiation, and tumor development.