LNCaP cells are human prostatic cancer cells that have a frame-shift mutation of the tumor suppressor gene PTEN and do not express the insulin receptor substrate-1 (IRS-1), a major substrate of the type 1 insulin-like growth factor receptor (IGF-IR). Ectopic expression of IRS-1 in LNCaP cells increases cell adhesion and decreases cell motility by an IGF-I-independent mechanism. We show now that these effects of IRS-1 are accompanied by serine phosphorylation of IRS-1 and are inhibited by inhibitors of phosphatidylinositol 3-kinase (PI3K). We have confirmed the requirement for PI3K activity and serine phosphorylation by the use of IRS-1 mutants, expressed in LNCaP cells. Serine phosphorylation inhibits IGF-I-induced tyrosyl phosphorylation of IRS-1, which is restored by the expression of wild-type PTEN or by inhibition of PI3K activity. Finally, IRS-1 in LNCaP cells co-immunoprecipitates with integrin α 5 β 1, and the association is again IGF-I-independent. We conclude that in LNCaP cells, IRS-1 is serine phosphorylated by PI3K, generating effects that are different, and even opposite, from those generated by IGF-I.
An attractive paradigm for studying how growth factors regulate the delicate balance between cell reproduction and cell death is offered by the insulin-like growth factor I receptor (IGF-IR), activated by its ligands. The IGF-IR is mitogenic in vivo and in vitro, promotes growth in size of the cell, sends a powerful anti-apoptotic signal, can induce differentiation in some cell types and plays a major role in the establishment and maintenance of the transformed phenotype. These various aspects of IGF-IR action have been discussed in detail in recent reviews by Blakesley et al. (1999); Baserga et al. (1999) and Grimberg and Cohen (2000). A less studied function of the IGF system is the ability of IGF-I to increase cell attachment and to alter cell motility (Jones et al., 1996; Doerr and Jones, 1996; Guvakova and Surmacz, 1997; Valentinis et al., 1998; Zheng and Clemmons, 1998; Dunn et al., 1998; Clemmons et al., 1999).
LNCaP cells are a human prostatic cancer cell line that originated from a metastatic tumor. Although they respond to IGF-I (Reiss et al., 1998, 2000), LNCaP cells have very low levels of IGF-IR and do not express IRS-1 (Reiss et al., 2000), a major substrate of the IGF-IR (White, 1998), and a strong activator of PI3K (Myers et al., 1994). The attenuation of the PI3K pathway is seemingly compensated in LNCaP cells by a frame-shift mutation of PTEN, a tumor suppressor gene (Li et al., 1997; Steck et al., 1997; Li and Sun, 1998) that regulates the activity of PI3K (Furnari et al., 1998; Li et al., 1998; Tamura et al., 1999; Davies et al., 1999). Like the IGF-IR, IRS-1 is also involved in the organization of the cytoskeleton (increased cell attachment and cell-to-cell aggregation, and altered cell motility), albeit in a somewhat different way. IGF-I increases both cell adhesion and cell motility (Doerr and Jones, 1996; Jones et al., 1996; Dunn et al., 1998; Zheng and Clemmons, 1998). Ectopic expression of IRS-1 in LNCaP cells increases cell attachment but markedly reduces cell motility (Reiss et al., 2000). The extinction of IRS-1 expression in LNCaP cells has been interpreted by Reiss et al. (2000) as a mechanism by which prostatic cancer cells could favor their metastatic spread by decreasing cell adhesion and increasing cell motility. The loss of the mitogenic stimulus of IRS-1 (D'Ambrosio et al., 1995; Tanaka et al., 1996; Peruzzi et al., 1999; Valentinis et al. 2000) is balanced by the PTEN mutation, which restores the activity of the PI3K pathway. Indeed, Akt, a downstream target of PI3K (reviewed in Chan et al., 1999), is constitutively activated in parental LNCaP cells (see Discussion).
We have explored the mechanisms by which IRS-1 regulates cell adhesion and cell motility in LNCaP cells. Both of these processes are complex processes, involving attachment to diverse substrates as well as cell-to-cell aggregation, and changes in cytoskeleton organization. However, for simplicity, we will refer to the effects of IRS-1 as effects on cell adhesion (including cell aggregation) and cell motility. We find that the effect of IRS-1 on cell adhesion and motility in LNCaP cells has some unusual aspects: (1) it requires serine phosphorylation of IRS-1 and the activation of PI3K; (2) it seems to require an interaction of IRS-1 with integrin α5β1; and (3) it is IGF-I-independent.
Phosphorylation of IRS-1 in LNCaP cells
In a previous study (Reiss et al., 2000), we reported that ectopic expression of IRS-1 in LNCaP cells increased cell adhesion and decreased cell motility. We also showed that over-expression of the IGF-IR in parental LNCaP cells caused growth arrest, while the combined expression of both the IGF-IR and IRS-1 resulted in optimal growth. We have explored IRS-1 phosphorylation in these cells, specifically we asked whether IRS-1 was tyrosyl phosphorylated or serine phosphorylated. The results of repeated experiments are summarized in Figure 1a showing a Western blot with an antibody to phosphorylated serine 612 of IRS-1. Serine 612 is among the IRS-1 serines involved in modulation of IRS-1 signaling (Mothe and van Obberghen, 1996). Lysates from LNCaP/IRS-1 cells (lane 1 of Figure 1a) show that wild-type IRS-1 is strongly positive for S612 phosphorylation. The controls were LNCaP transduced with the empty vector (lane 2, no detectable band) and R600 cells (lane 3), where a very faint band is detectable. These experiments were carried out in serum-free medium, which may explain the very faint band in lysates from R600 cells. IRS-1 is unambiguously tyrosyl phosphorylated when R600 cells are stimulated with IGF-I (Rubini et al., 1997).
Since PI3K is known to be a serine kinase that can serine phosphorylate IRS-1 (Lam et al., 1994; Freund et al., 1995; Cengel and Freund, 1999), we formulated the hypothesis that serine phosphorylation of IRS-1 could be due to PI3K activity. For this purpose, we examined the effect of an inhibitor of PI3K, LY294002 on serine phosphorylation of IRS-1. A convenient way to detect serine phosphorylation of IRS-1 is by mobility shift (Mothe and van Obberghen, 1996; D'Ambrosio et al., 1997). Figure 1b shows that LY294002 increases the mobility of IRS-1, confirming that in these cells IRS-1 is serine phosphorylated and suggesting that this phosphorylation is dependent on PI3K activity (see also below). Addition of IGF-I has no appreciable effect on the mobility of IRS-1.
It is generally agreed that serine phosphorylation and tyrosine phosphorylation of IRS-1 are, at least to a certain extent, mutually exclusive (Feinstein et al., 1993; Kanety et al., 1995). We therefore examined tyrosyl phosphorylation of IRS-1. In LNCaP/IRS-1 cells, IGF-I induces a barely detectable tyrosyl phosphorylation of IRS-1 (Figure 1c, lane 2), which is not detectable in LNCaP cells transduced with the empty vector (lane 4). However, a strong band appears in the same cells treated with LY294002 (Figure 1c, lane 6), indicating that this inhibitor reverses the serine phosphorylation of IRS-1 and restores the ability of IGF-I to induce tyrosyl phosphorylation of its substrate. Tyrosyl phosphorylation of IRS-1 is also clearly visible (Figure 1c, lane 8) in LNCaP/IRS-1 cells transduced with a wild-type PTEN (Reiss et al., 2000). In these two last conditions, tyrosyl phosphorylation of IRS-1 leads to a commensurate decrease in positivity to S612 antibody (data not shown). These experiments indicate that wild-type IRS-1 is serine phosphorylated in LNCaP/IRS-1 cells, and that this serine phosphorylation is reduced by inhibitors of the PI3K pathway.
Effect of inhibitors of PI3-kinase on IRS-1-mediated increase in cell aggregation
We previously reported that IRS-1 expression in LNCaP cells increases cell-to-cell aggregation (Reiss et al., 2000). The extent of cell aggregation can be semi-quantified by measuring the size and the number of the aggregates formed when cells are seeded on polyHEMA-coated plates (Valentinis et al., 1998; Reiss et al., 2000). Since the experiments detailed above suggested that PI3K may play a crucial role in the effect of IRS-1 on cell adhesion in LNCaP cells, we tested this possibility by asking whether an inhibitor of PI3K could inhibit cell-to-cell aggregation. The inhibitor LY294002 was tested on LNCaP cells expressing the wild-type IRS-1 in either the presence or absence of IGF-I. In this assay, we measured the sizes of the aggregates and determined the number of such aggregates within certain size ranges (Figure 2). The addition of IGF-I increases cell aggregation (Guvakova and Surmacz, 1997; Valentinis et al., 1998) of parental LNCaP cells (actually LNCaP cells transduced with an empty vector). The number of aggregates in the 300–600 μM range increased from eight in serum-free medium to 29 in IGF-I (Figure 2, upper panel). The size of the aggregates increased sharply in LNCaP/IRS-1 cells (lower panel). For instance, taking the results obtained with the cells in IGF-I, the number of aggregates with a size range between 300 and 600 μM increased from 29 to 41. In addition, 17 aggregates were detected in LNCaP/IRS-1 cells in the size range 601–1200 μM, against none in the parental cells. LY294002 markedly decreases the number of the larger aggregates, especially evident in LNCaP/IRS-1 cells in IGF-I (Figure 2). These results further support a role of PI3K in the effect of IRS-1 on cell adhesion in LNCaP cells.
Generation of IRS-1 mutants and cell lines
We have confirmed the role of PI3K in IRS-1-mediated cell adhesion and motility by using a set of IRS-1 mutants. Figure 3a, is a diagram of IRS-1 and of the mutations used in these studies. Similar mutant IRS-1 have already been used to study insulin signaling in 32D cells (Yenush et al., 1998), so that their short term effects on insulin signaling are known. In those studies, cell attachment and motility were not included in the analysis. Briefly, the Pleckstrin domain (PH) and the phosphotyrosine binding domain (PTB) contribute strongly to the interaction of IRS-1 with the insulin or IGF-I receptors. The tyrosine motifs indicated in Figure 3a, recognize the p85 subunit of PI3-kinase or Grb2 (reviewed in White, 1998 and Ogawa et al., 1998). In 32D cells, all these mutants are at least partially functional, including the PH/PTB mutant (Yenush et al., 1998), that includes only the first 309 residues of the IRS-1 protein, and the δ PTB mutant (Yenush et al., 1996). Our IRS-1 residue numbers are slightly different from those of Ogawa et al. (1998), because the IRS-1 we used is of mouse origin (D'Ambrosio et al., 1995). The tyrosine residue in the Grb2 binding site is at position 891 (instead of 895), and the tyrosine in the distal binding site for p85 is at position 935 (instead of 939).
Figure 3b, gives the levels of expression of the IRS-1 mutants transduced into LNCaP cells. All of these cell lines are mixed populations, resulting from retroviral transduction. The PH/PTB mutant could only be detected with an antibody to the amino-terminus of IRS-1 (lower half of Figure 3b). The other mutants can be detected with an antibody to the carboxy-terminus (upper half of Figure 3b). The mutants with deletions of either the PTB or the PH domains migrate, as expected, faster than wild-type IRS-1 or the point mutants. We could not obtain, despite repeated attempts, a LNCaP cell line that would give a strong expression of the Grb2 mutant (single point mutation). We suspect that the Grb2 binding site may contribute to the stability of IRS-1, but this point was not further pursued in these studies. All other mutants are strongly expressed in LNCaP cells, as is the wild-type IRS-1. Figure 3b, also confirms our previous observations that parental LNCaP cells do not express a detectable IRS-1.
Effect of IRS-1 mutations on its ability to decrease cell motility
Wild-type IRS-1 was previously shown to markedly decrease cell motility of LNCaP cells as tested in a Boyden chamber (Reiss et al., 2000). We now tested cell motility in LNCaP cells stably expressing mutant IRS-1 proteins, using the cell lines described above. Ectopic expression of IRS-1 markedly decreases motility in LNCaP cells, by about 70% (Figure 4), even when the cells are in serum-free medium, as in the previous paper (Reiss et al., 2000). IGF-I increases cell motility in parental (empty vector) LNCaP cells. An IGF-I-mediated increase in cell motility has been repeatedly reported in the literature (Doerr and Jones, 1996; Dunn et al., 1998; Zheng and Clemmons, 1998). There is therefore a discrepancy in the effects of IGF-I and IRS-1 on cell motility. Since IGF-I still increases cell motility in parental cells, the seemingly paradoxical effect of IRS-1 on cell motility in LNCaP cells suggests that it may be IGF-I-independent (see also below).
As to the mutant IRS-1 (Figure 4, lower panel), only the mutant IRS-1 with a deletion of the PH domain significantly reduces cell motility (45 against 70% with the wild-type). All other mutants lose this function, although some of the mutants show a slight decrease, around 20%, but with some variability. The inability of the mutant at residues Y608 and Y939 (PI3K binding sites) to sustain the action of IRS-1 in LNCaP cells again suggests that PI3K may be required for the inhibition of cell motility. This is further supported by the inability of the PH/PTB mutant (also lacking the PI3K binding sites) and of the δPTB mutant to decrease cell motility. A mutant IRS-1 with a deletion of the PTB domain is tyrosyl phosphorylated by the insulin receptor, but fails to stimulate PI3K activity (Yenush et al., 1996).
Effect of IRS-1 mutations on cell attachment
Cell attachment was determined by the method previously described (Dunn et al., 1998; Reiss et al., 2000). We confirm that ectopic expression of IRS-1 increases attachment of LNCaP cells to fibronectin, laminin and collagen I (Figure 5), whether the cells are tested in serum-free medium (SFM) or in SFM supplemented with IGF-I (shown only for laminin). The increase, as in a previous paper (Reiss et al., 2000), is not dramatic (only a doubling of the percentage of cells attached to a substratum), but it is significant and it is highly reproducible. IGF-I, by itself, increases cell attachment in parental as well as in IRS-1 expressing LNCaP cells. As to the mutant IRS-1, the results (Figure 5), repeated several times, were consistent. All IRS-1 mutants, with one exception, lose the ability to increase the attachment of LNCaP cells to the substrata. The exception is again the mutant with a deletion of the PH domain. Both on laminin and fibronectin, the δPH mutant is not as effective as the wild-type IRS-1, but it does significantly increase cell attachment. It is essentially as effective as the wild-type IRS-1 when the cells are tested on collagen I.
Interaction of IRS-1 with α5β1 integrin
Integrins are known to be involved in IGF-I signaling, especially in cell motility (see Discussion). The integrin that is usually associated with the IGF axis is the αVβ3, which has been reported to interact directly with IRS-1 (Vuori and Ruoslahti, 1994). We were not able to co-precipitate IRS-1 and αVβ3 in LNCaP cells, but IRS-1 was co-precipitated by an antibody to integrin α5β1 (Figure 6a) shows that IRS-1 co-precipitation occurs only when IRS-1 is active in cell adhesion and motility, i.e. with the wild-type IRS-1 and the Pleckstrin deletion mutant. There is no co-precipitation with the inactive IRS-1 mutants. The co-precipitation occurs in serum-free medium, and it is not augmented by stimulation with IGF-I (Figure 6b), confirming that the effect of IRS-1 on LNCaP cells can be IGF-I-independent.
We have used the same mutants to explore further the serine phosphorylation of IRS-1 in LNCaP cells (see Figure 1). By mobility shift, serine phosphorylation is undetectable in IRS-1 mutants that have lost their ability to increase cell adhesion and decrease cell motility. Figure 6c, shows that wild-type IRS-1 (lane 3) shows an obvious mobility shift, when compared with the two mixed populations, expressing two mutants of IRS-1, that have lost the ability of affecting the cytoskeleton organization of LNCaP cells (lanes 1 and 2). These two mutants of IRS-1 are the 2 point and 3 point mutants. It is known that serine phosphorylation of IRS-1 decreases its mobility (Mothe and van Obberghen, 1996; D'Ambrosio et al., 1997). When the lysates from LNCaP/IRS-1 cells were treated with alkaline phosphatase, that removes phosphates from serines (Ceresa and Pessin, 1996), the mobility shift is abolished (lane 4), and the IRS-1 band is at the same level as the bands of the two inactive mutants. Figure 6c, lanes 5 and 6 show the wild-type IRS-1 in cells treated with okadaic acid (OKA, lane 5), and the same lysate after treatment with alkaline phosphatase (lane 6). Finally, Figure 6c, lane 7 shows that the mobility of wild-type IRS-1 returns to normal when LNCaP/IRS-1 cells are treated with LY294002. The mobility shift with OKA is more pronounced than in untreated cells, but OKA is a powerful inhibitor of serine/threonine phosphatases, and probably results in a massive serine phosphorylation of IRS-1 (Mothe and van Obberghen, 1996).
The independence from IGF-I in these effects of IRS-1 on prostate cancer cells is one of the most interesting findings in these experiments. It could be objected that LNCaP cells may produce IGF-I and grow by an autocrine mechanism; they do indeed grow slightly in serum-free medium, although they are stimulated by IGF-I (Reiss et al., 1998). To confirm the IGF-I independence, we used R-cells, which are mouse embryo fibroblasts with a targeted disruption of the IGF-I genes (Sell et al., 1993). These cells have been studied extensively in our laboratory, they are totally devoid of IGF-IR and they do not respond to IGF-I. Figure 6d shows that the interaction between IRS-1 and the α5β1 integrin occurs also in R-cells.
Effect of IRS-1 on the morphology of LNCaP cells
Since wild-type IRS-1 seems to have a profound effect on cell adhesion and motility in LNCaP cells, we have asked whether its expression could also change the morphology of these cells. Figure 7 shows that the morphology is indeed changed. Since these are epithelial cells, actin cables are not as evident as in fibroblasts. However, LNCaP/IRS-1 cells are clearly different from parental cells. They are more rounded (Figure 7d), and they have a tendency to grow in clusters, even in monolayer cultures. This is compatible with previous reports that PI3K and its products induce cytoskeletal reorganization (Rodriguez-Viciana et al., 1997; Raucher et al., 2000).
In a previous study (Reiss et al., 2000), we reported that expression of IRS-1 in LNCaP cells increased cell adhesion and decreased cell motility, two effects that could play a role in favoring the metastatic spread of prostate cancer cells. We show now that both cell adhesion (cell attachment and cell-to-cell aggregation) and cell motility changes are dependent on serine phosphorylation of IRS-1 and a functional PI3K. Specifically, the novel findings include: (1) IRS-1 is serine phosphorylated in LNCap/IRS-1 cells, even when the cells are stimulated with IGF-I; (2) serine phosphorylation and cell attachment are inhibited by inhibitors of PI3K; (3) mutant IRS-1 defective in PI3K signaling are also defective in increasing cell attachment and decreasing cell motility; (4) active IRS-1 interacts directly with α5β1 integrin; and (5) all these effects are IGF-I independent.
We would like to begin the discussion by noting that LNCaP cells do respond to IGF-I, despite the low number of receptors. The ability of LNCaP cells to respond to IGF-I was documented in previous reports (Pietrzkowski et al., 1993; Reiss et al., 1998), and is confirmed in this paper. IGF-I increases cell adhesion to laminin (Figure 5), and increases cell motility (Figure 4) of parental LNCaP cells. These are the same results obtained by other investigators with different cell types (Doerr and Jones, 1996; Dunn et al., 1998; Zheng and Clemmons, 1998). The first important observation is that IRS-1 dissociates the two processes, as ectopic expression of IRS-1 in LNCaP cells increases cell attachment but decreases cell motility (Reiss et al., 2000 and this paper). The second important observation is that the effect of IRS-1 is IGF-I-independent, which may explain why IGF-I and IRS-1 are not concordant in their effects.
We have now asked how IRS-1 increases cell adhesion and cell aggregation and decreases cell motility in prostate cancer cells. In the previous paper, we reported that ectopic expression of IRS-1 in LNCaP cells results in a dramatic increase in PI3K activity (Reiss et al., 2000). However, paradoxically, there is very little tyrosyl phosphorylation of IRS-1, after IGF-I stimulation (Figure 1c, lane 2). This could be due to the fact that, under these conditions, IRS-1 is serine phosphorylated, as indicated both by an antibody specific for phosphorylation of serine 612, a serine involved in IRS-1 signaling (Mothe and van Obberghen, 1996) and by mobility shift. PI3K is a serine kinase (Lam et al., 1994; Freund et al., 1995; Delahaye et al., 1998; Cengel and Freund, 1999), and there are several reports that, under certain circumstances, IRS-1 is serine phosphorylated (for references, see D'Ambrosio et al., 1997). In general, serine phosphorylation of IRS-1 reduces its tyrosyl phosphorylation (Mothe and van Obberghen, 1996; Cengel and Freund, 1999) and signaling (Feinstein et al., 1993; Kanety et al., 1995; Staubs et al., 1998). We have confirmed these observations, since IRS-1, when expressed in LNCaP cells, is largely serine phosphorylated, with a very modest extent of tyrosyl phosphorylation. Lam et al. (1994) have already reported that inhibitors of PI3K inhibit PI3K serine kinase activity. If PI3K is involved in serine phosphorylation of IRS-1 and its effects on cell adhesion, the PI3K inhibitor LY294002 should reduce IRS-1 serine phosphorylation, increase its tyrosyl phosphorylation and inhibit its effect on cell aggregation. This is exactly what we have observed in these experiments. To obtain other clues about the mechanism of IRS-1-mediated effects on LNCaP cells, we have transduced parental cells with IRS-1 expression vectors, carrying diverse IRS-1 mutations (Figure 3). All the mutations we have made, with one exception, abrogate the ability of IRS-1 to alter cell adhesion and motility. The exception is the deletion of the Pleckstrin domain (δPH). Three of the mutants (2 point, 3 point mutations and PH/PTB) have been mutated in or are missing the binding sites for PI3K (Ogawa et al., 1998). Their inability to alter cell adhesion and motility in LNCaP cells confirms that PI3K is involved in the effect of IRS-1 on these processes. Interestingly, these inactive mutants also fail to be serine phosphorylated, as indicated in Figure 6.
There is actually a small but reproducible impairment of IRS-1 effect, when the Pleckstrin domain is deleted. Still, the δPH mutant does increase cell attachment and decreases cell motility. This impairment may be related to the fact that the Pleckstrin domain may be important for binding of IRS-1 to the receptor. It is so in the case of the insulin receptor, especially when the receptor levels are low. With over-expression of the insulin receptor, the PTB domain also becomes involved in the interaction with the receptor (Yenush et al., 1996). In 32D cells over-expressing the insulin receptor, none of these mutant IRS-1 is a completely disabled IRS-1. These mutants may be defective in some functions, but generally, they transmit signals, including the ability to induce thymidine incorporation into cells (Yenush et al., 1996, 1998). The results so far can be interpreted as follows: in LNCaP/IRS-1 cells, increase in cell adhesion and decrease in cell motility are mediated by the activation of PI3K. The activation of PI3K, instead of sending the customary signal to the Akt/PKB pathway, results in serine phosphorylation of IRS-1. The next question we asked was why PI3-kinase would focus on serine phosphorylation of IRS-1, instead of Akt. A reasonable explanation is that in LNCaP cells, the Akt/PKB pathway is already hyper-activated. Akt is constitutively activated in parental LNCaP cells (Davies et al., 1999; Carson et al., 1999; Tamura et al., 1999; Reiss et al., 2000). Akt/PKB and IRS-1 form stable complexes in vivo, and over-expression of Akt/PKB enhances serine phosphorylation of IRS-1 (Li et al., 1999; Paz et al., 1999). This interpretation is supported by the fact that introduction into LNCaP/IRS-1 cells of a wild-type PTEN, which abrogates the constitutive activation of Akt/PKB (Reiss et al., 2000), results in tyrosyl phosphorylation of IRS-1. Incidentally, the effect of IRS-1 on cell adhesion is not limited to LNCaP cells. We have observed that ectopic expression of IRS-1 in 32D cells, a murine hemopoietic cell line that normally grows in suspension, causes these cells to adhere to the plate. The strength of the attachment is proportional to the levels of expression of IRS-1 (unpublished data from our laboratory).
The last question is how IRS-1 function is activated in this model system. At this point, we can only offer some conjectures. We have examined LNCaP cells in monolayers, and there are reports that IRS-1 interacts with integrins (Guilherme and Czech, 1998), and that there is cross-talk between the IGF axis and integrins, especially αVβ3 (Vuori and Ruoslahti, 1994; Jones et al., 1996; Brooks et al., 1997; Clemmons et al., 1999). We have confirmed an interaction between IRS-1 and integrins, although, in our cells, it is the α5β1 integrin that co-precipitates IRS-1. The question may be raised why an interaction with α5β1 integrin affects cell adhesion not only with fibronectin but also with laminin and collagen 1. The reasonable answer is that IRS-1 interacts also with αVβ3 (see references above), which, in prostate cancer cells is expressed at much lower levels than α5β1 integrin (Witkowski et al., 1993; Zheng et al., 1999). Because of the scarcity of αVβ3 in prostate cancer cells, we could not detect a clear interaction of IRS-1 with this integrin. It may still be possible that there is enough αVβ3 integrin in our cells to give a biological effect with IRS-1. The IRS-1/integrin interaction is again IGF-I-independent. The independence of this interaction from IGF-I has been rigorously confirmed by showing that the interaction occurs also in R-cells, totally devoid of IGF-IR and totally insensitive to IGF-I (Sell et al., 1993). IRS-1 is also known to be activated, when cells are plated on appropriate substrates (Zheng and Clemmons, 1998; Guilherme et al., 1998), which brings up the possibility of FAK involvement (Valentinis et al., 1998). Indeed, IRS-1 has been reported to be a signaling molecule for pp125FAK (Lebrun et al., 1998). IRS-1 has a Grb2 binding site, and Grb2 is known to interact with integrins (Shakibaei et al., 1999) and integrins with FAK (Renshaw et al., 1999). This could explain why the Grb2 mutant in our experiments is also inactive. Alternatively, it could be PI3K itself that is activated by interaction with integrins (Guilherme and Czech, 1998), leading to serine phosphorylation of IRS-1.
In conclusion, the picture that is emerging from these studies is the following. Ectopic expression of IRS-1 in LNCaP cells results in its phosphorylation at serine residues, perhaps caused by the constitutive activation of Akt/PKB (because of the PTEN mutation). Serine phosphorylation is known to cause degradation of IRS-1 (Sarbassov and Peterson, 1998), offering a possible mechanism for the extinction of IRS-1 expression in these cells. Incidentally, IRS-1 degradation is inhibited by inhibitors of PI3K (Lee et al., 2000). The serine phosphorylation of IRS-1 in this model requires the activation of PI3K. The effect of IRS-1 on cell adhesion and cell motility is somewhat unique, since it does not require IGF-I. In fact, IRS-1 and IGF-I have opposite effects on cell motility. IRS-1 signaling under these conditions is probably due to its activation by integrins (in our cells the α5β1 integrin) or other components of the extra-cellular matrix system. The possibility that extinction of IRS-1 expression in LNCaP cells may have a profound effect on the metastatic spread of human prostatic cancer has already been discussed by Reiss et al. (2000). The extinction of IRS-1 expression combined with a PTEN mutation could facilitate the metastatic spread of cells, without impairment of the strongly mitogenic Akt/PKB pathway.
Materials and methods
Plasmid pGR157 is based on a self-inactivating form of the MSCV retroviral vector system (Hawley et al., 1994) and contains an internal CMV promoter derived from plasmid pcDNA3.1(+) (Invitrogen, Carlsbad, CA, USA). The self-inactivating form of MSCV retroviral vector was generated as described elsewhere (Yu et al., 1986). Plasmid pGR159 was generated by inserting the wild-type mouse IRS-1 sequence into the HindIII restriction site of plasmid pGR157. This wild-type mouse IRS-1 lacks the 3′ untranslated region and is under the control of the CMV promoter. All the mutants of IRS-1 were generated from plasmid pGR159 by standard procedures, using mutagenesis kits (Stratagene, La Jolla, CA, USA). The PCR reactions and the purification of the PCR products have been previously described (Romano et al., 1999).
Plasmid pGR171 contains a truncated form of mouse IRS-1. This truncated mouse IRS-1 comprises the first 309 aminoacids, which include the pleckstrin domain (PH) and the phospho-tyrosine binding domain (PTB). Plasmid PGR163 contains a mouse IRS-1 lacking the PTB region (from amino acid 155 to 309). Plasmid pGR167 contains a mouse IRS-1 gene lacking the PH region (first 107 amino acids from the start codon). Plasmid pGR169 contains a point mutated mouse IRS-1 gene. The point mutation changed a tyrosine residue at position 891 into phenylalanine (Y891F). Plasmid pGR181 contains a double point mutated mouse IRS-1 gene. The point mutations changed two tyrosine residues at positions 608 and 935 into phenylalanine (Y608F+Y935F). Plasmid pGR194 contains a triple mutated mouse IRS-1 gene. The point mutations changed three tyrosine residues at positions 608, 891 and 935 into phenylalanine (Y608F+Y891F+Y935F). The sequences of all the mutations were monitored for the presence of the specific mutations and for possible misincorporations that could have been accidentally introduced.
Cells and cell culture
Parental LNCaP cells, human metastatic prostate adenocarcinoma cells (ATCC CRL-1740), were cultured in RPMI (GIBCO–BRL, Grand Island, NY, USA) supplemented with 50 U/ml penicillin, 50 μg/ml streptomycin (P/S), and 10% fetal bovine serum (FBS) (SIGMA, St. Louis, MO, USA). Mixed populations of LNCaP cells expressing wt mouse IRS-1 cDNA (LNCaP/IRS-1) or IRS-1 mutants (see above) were generated by retroviral transductions according to the protocol previously described (Romano et al., 1999). LNCaP/IRS-1 cells expressing ectopic PTEN or IGF-IR cDNAs were previously described (Reiss et al., 2000).
Quiescent cultures were prepared by plating 2.5×103 cells/cm2 in the presence of 10% FBS. After 24 h cells were washed three times with Hank's balanced salt solution (HBSS) and cultured in a serum-free medium (SFM) [RPMI supplemented with 0.1% bovine serum albumin (SIGMA) and glutamine (GIBCO)]. Generally, LNCaP cells become quiescent after 48 h in SFM, which was renewed with fresh medium every 24 h. To test effects of growth factors and metabolic inhibitors on cell motility, attachment, and aggregation, or IRS-1 phosphorylation, quiescent cells were pretreated for 1 h with LY294002 (10 or 50 μM), or for 40 min with okadaic acid (2 μM), prior to IGF-I (50 ng/ml) stimulation.
To determine the morphology of F-actin filaments, 24 h after plating, cultures of LNCaP and LNCaP/IRS-1 cells were pretreated with 0.1% Triton X-100 for 5 min and fixed with 3.7% formaldehyde. The cells were then washed three times in PBS and incubated in the PBS solution containing 1% BSA and 1 U of Texas Red–X phalloidin as recommended by manufacturer's protocol (Molecular Probes, Inc., Eugene, Oregon, USA).
Immunoprecipitation and immunobloting
These were performed according to methodologies repeatedly used in our previous work (Valentinis et al., 1998; Peruzzi et al., 1999; Reiss et al., 2000). Briefly, cells were lysed on ice with a 400 μl of lysis buffer [50 mM N-2-hydroxyethylpiperazine N′-2-ethanesulfonic acid (HEPES), pH 7.5, 150 mM NaCl, 1.5 mM MgCl2, 1 mM EGTA, 10% glycerol, 1% Triton X-100, 1% phenylmethylsulfonyl fluoride (PMSF), 0.2 mM Na-orthovanadate and 1% aprotinin], and protein concentration was determined by a Bio-Rad Protein Assay (BioRad, Hercules, CA, USA). A 500 μg of total protein extract was used for immunoprecipitation. IRS-1 was immunoprecipitated with anti-IRS-1 polyclonal antibody (Upstate Biotechnology, Lake Placid, NY, USA), and IRS-1 tyrosine phosphorylation detected with anti PY20-HRP conjugated antibody (Transduction Laboratories). Western blots for wt IRS-1 and most of the IRS-1 mutants were done with an anti-IRS-1 antibody against the C-terminal 14 amino acids (Upstate Biotechnology). Anti-IRS-1 N-terminal antibody (Santa Cruz) was utilized to detect the PH/PTB mutant of IRS-1. Serine phosphorylation of IRS-1 was evaluated by retardation in gel mobility (Mothe and van Obberghen, 1996), and by utilizing anti-IRS-1 (S612) phosphospecific antibody (BioSource International, Camarillo, CA, USA). To dephosphorylate IRS-1, immunoprecipitated samples were treated with Alkaline Phosphatase (SIGMA; 1000 U) for 60 min at 37°C, as previously described (Ceresa and Pessin, 1996) and the dephosphorylated samples were subjected to gel electrophoresis on 4–15% gradient SDS–PAGE (BioRad). In one instance (Figure 6c), the proteins were resolved on a 7.5% PAGE, in order to detect more clearly the mobility shifts.
Attachment and motility assays
The effects of ectopic IRS-1 expression and its mutants on cell adhesion to different extra-cellular matrix (ECM) proteins were evaluated accordingly to methodology previously described (Dunn et al., 1998) with some modifications (Reiss et al., 2000). For the aggregation assay, we followed the protocol described in our previous work (Valentinis et al., 1998). Evaluation of size and number of aggregates was done under the inverted phase-contrast microscope. Cell motility was assessed by utilizing modified Boyden Chambers (Bicoat Inserts, Becton-Dickinson) as previously described (Guvakova and Surmacz, 1997). Fibronectin-coated inserts containing polyethylene terephthalate (PTE) membranes with 8 μm pores at 1×105 pores/cm2 were employed to evaluate cell migration (Reiss et al., 2000).
To determine interaction between IRS-1 and α5β1 integrin, 500 μg of total proteins were incubated with 10 μg of anti-human-α5β1 mouse antibody (MAB1969; Chemicon International, Inc., Temecula, CA, USA) and 20 μl of anti-mouse agarose-conjugated IgG (Sigma) in a total volume of 500 μl of HNTG buffer. Following overnight incubation at 4°C,immuno-complexes were washed 3 Times in HNTG buffer, diluted in 20 μl of Laemmli sample buffer (BioRad), boiled for 5 min and separated on 4–15% gradient SDS–PAGE (BioRad). Following transfer, nitrocellulose membranes were probed with anti-IRS-1 rabbit antibody (Upstate Biotechnology). The same technique was followed for R-cells, except that the antibody anti-α5β1 integrin was against the mouse protein.
Baserga R, Prisco M and Hongo A. . 1999 IGFs and Cell Growth. In: The IGF System.Rosenfeld R and Roberts Jr C. (eds). Humana Press: Totowa, NJ pp. 329–353.
Blakesley VA, Butler AA, Koval AP, Okubo Y and LeRoith D. . 1999 IGF-I receptor function: transducing the IGF-I signal into intracellular events. The IGF System. Rosenfeld R and Roberts Jr C. (eds). Humana Press: Totowa, NJ pp. 143–163.
Brooks PC, Klemke RL, Schoen S, Lewis JM, Schwartz MA and Cheresh DA. . 1997 J. Clin. Invest. 99: 1390–1398.
Carson JP, Kulik G and Weber MJ. . 1999 Cancer Res. 59: 1449–1453.
Cengel KA and Freund GG. . 1999 J. Biol. Chem. 274: 27969–27974.
Ceresa BP and Pessin JE. . 1996 J. Biol. Chem. 271: 12121–12124.
Chan TO, Rittenhouse SE and Tsichlis PN. . 1999 Ann. Rev. Biochem. 68: 965–1014.
Clemmons DR, Horvitz G, Engleman W, Nichols T, Moralez A and Nickols GA. . 1999 Endocrinology 140: 4616–4621.
D'Ambrosio C, Valentinis B, Prisco M, Reiss K, Rubini M and Baserga R. . 1997 Cancer Res. 57: 3264–3271.
D'Ambrosio C, Keller SR, Morrione A, Lienhard GE, Baserga R and Surmacz E. . 1995 Cell Growth Diff. 6: 557–562.
Davies MA, Koul D, Dhesi H, Berman R, McDonnell TJ, McConkey D, Yung WKA and Steck PA. . 1999 Cancer Res. 59: 2551–2556.
Delahaye L, Mothe-Satney I, Myers MG, White MF and van Obberghen E. . 1998 Endocrinology 139: 4911–4919.
Doerr ME and Jones JL. . 1996 J. Biol. Chem. 271: 2443–2447.
Dunn SE, Ehrlich M, Sharp NJH, Reiss K, Solomon G, Hawkins R, Baserga R and Barrett JC. . 1998 Cancer Res. 58: 3353–3361.
Feinstein R, Kanety H, Papa MZ, Lunenfeld R and Karasik A. . 1993 J. Biol. Chem. 268: 26055–26058.
Freund GG, Wittig JG and Mooney RA. . 1995 Biochem. Biophys. Res. Comm. 206: 272–278.
Furnari FB, Huang HJS and Cavenee WK. . 1998 Cancer Res. 58: 5002–5008.
Grimberg A and Cohen P. . 2000 J. Cell. Physiol. 183:: 1–9.
Guilherme A and Czech MP. . 1998 J. Biol. Chem. 273: 33119–33122.
Guilherme A, Torres K and Czech MP. . 1998 J. Biol. Chem. 273: 22899–22903.
Guvakova MA and Surmacz E. . 1997 Exp. Cell Res. 231: 149–162.
Hawley RG, Lieu FHL, Fong AZC and Hawley TS. . 1994 Gene Ther. 1: 136–138.
Jones JI, Prevette T, Cockerman A and Clemmons DR. . 1996 Proc. Natl. Acad. Sci. USA 93: 2482–2487.
Kanety H, Feinstein R, Papa MZ, Hemi R and Karasik A. . 1995 J. Biol. Chem. 270: 23780–23784.
Lam K, Carpenter CL, Ruderman NB, Friel JC and Kelly KL. . 1994 J. Biol. Chem. 269: 20648–20652.
Lebrun P, Mothe-Satney I, Delahaye L, Van Obberghen E and Baron V. . 1998 J. Biol. Chem. 273: 32244–32253.
Lee AV, Gooch JL, Oesterreich S, Guler RL and Yee D. . 2000 Mol. Cell. Biol. 20: 1489–1496.
Li DM and Sun H. . 1998 Proc. Natl. Acad. Sci. USA 95: 15406–15411.
Li J, Yen C, Liaw D, Podsypanina K, Bose S, Wang SI, Ouc J, Miliaresis C, Rodgers L, McCombie R, Bigner SH, Giovanella BC, Ittmann C, Tycko B, Hibshoosh H, Wigler MH and Parsons R. . 1997 Science 275: 1943–1947.
Li J, De Fea K and Roth RA. . 1999 J. Biol. Chem. 274: 9351–9356.
Li J, Simpson L, Takahashi M, Miliaresis C, Myers MP, Tonks N and Parsons R. . 1998 Cancer Res. 58: 5667–5672.
Mothe I and van Obberghen E. . 1996 J. Biol. Chem. 271: 11222–11227.
Myers Jr MG, Grammer TC, Wang LM, Sun XJ, Pierce JH, Blenis J and White MF. . 1994 J. Biol. Chem. 269: 28783–28789.
Ogawa W, Matozaki T and Kasuga M. . 1998 Mol. Cell. Biochem. 182: 13–22.
Paz K, Liu YF, Shorer H, Hemi R, LeRoith D, Quan M, Kanety H, Seger R and Zick Y. . 1999 J. Biol. Chem. 274: 28816–28822.
Peruzzi F, Prisco M, Dews M, Salomoni P, Grassilli E, Romano G, Calabretta B and Baserga R. . 1999 Mol. Cell. Biol. 19: 7203–7215.
Pietrzkowski Z, Mulholland G, Gomella L, Jameson BA, Wernicke D and Baserga R. . 1993 Cancer Res. 53: 1102–1106.
Raucher D, Stauffer T, Chen W, Shen K, Guo S, York JD, Sheetz MP and Meyer T. . 2000 Cell 100: 221–228.
Reiss K, D'Ambrosio C, Tu X, Tu C and Baserga R. . 1998 Clin. Cancer Res. 4: 2647–2655.
Reiss K, Wang JY, Romano G, Furnari FB, Cavenee WK, Morrione A, Tu X and Baserga R. . 2000 Oncogene 12: 2687–2694.
Renshaw MW, Price LS and Schwartz MA. . 1999 J. Cell Biol. 147: 611–618.
Rodriguez-Viciana P, Warne pH, Khwaja A, Marte BM, Pappin D, Das P, Waterfield MD, Ridley A and Downward J. . 1997 Cell 89: 457–467.
Romano G, Prisco M, Zanocco-Marani T, Peruzzi F, Valentinis B and Baserga R. . 1999 J. Cell. Biochem. 72: 294–310.
Rubini M, Hongo A, D'Ambrosio C and Baserga R. . 1997 Exp. Cell Res. 230: 284–292.
Sarbassov DD and Peterson CA. . 1998 Mol. Endocrinol. 12: 1870–1878.
Sell C, Rubini M, Rubin R, Liu JP, Efstratiadis A and Baserga R. . 1993 Proc. Natl. Acad. Sci. USA 90: 11217–11221.
Shakibaei M, Joh T, de Souza P, Rhamanzadeh R and Merker HJ. . 1999 Biochem. J. 342: 615–623.
Staubs PA, Nelson JG, Reichart DR and Olefsky JM. . 1998 J. Biol. Chem. 273: 25139–25147.
Steck PA, Pershouse MA, Jasser SA, Yung WKA, Lin H, Ligon AH, Langford LA, Baumgard ML, Hattier T, Davis T, Frye C, Hu R, Swedlund B, Teng DHF and Tavtigian SV. . 1997 Nature Genetics 15: 356–362.
Tamura M, Gu J, Danen EHJ, Takino T, Miyamoto S and Yamada KM. . 1999 J. Biol. Chem. 274: 20693–20703.
Tanaka S, Ito T and Wands JR. . 1996 J. Biol. Chem. 271: 14610–14616.
Valentinis B, Romano G, Peruzzi F, Morrione A, Prisco M, Soddu S, Cristofanelli B, Sacchi A and Baserga R. . 2000 J. Biol. Chem. 274: 12423–12430.
Valentinis B, Reiss K and Baserga R. . 1998 J. Cell. Physiol. 176: 648–657.
Vuori K and Ruoslahti E. . 1994 Science 266: 1576–1578.
White MF. . 1998 Mol. Cell. Biochem. 182: 3–11.
Witkowski CM, Rabinovitz I, Nagle RB, Affinito KS and Cress AE. . 1993 Cancer Res. Clin. Oncol. 119: 637–644.
Yenush L, Zanella C, Uchida T, Bernal D and White MF. . 1998 Mol. Cell. Biol. 18: 6784–6794.
Yenush L, Makati KJ, Smith-Hall J, Ishibashi O, Myers Jr MG and White MF. . 1996 J. Biol. Chem. 271: 24300–24306.
Yu S-F., von Ruden T, Kantoff PW, Garber C, Seiberg M, Ruther U, Anderson WF, Wagner EF and Gilboa E. . 1986 Proc. Natl. Acad. Sci. USA 83: 3194–3198.
Zheng B and Clemmons DR. . 1998 Proc. Natl. Acad. Sci. USA 95: 11217–11222.
Zheng DQ, Woodard AS, Fornaro M, Tallini G and Languino LR. . 1999 Cancer Res. 59: 1655–1664.
This work is supported by grants CA 56309 and AG 16291 from the National Institutes of Health. K Reiss is a recipient of grant DHHS PO-1 NS 36466.
About this article
Cite this article
Reiss, K., Wang, J., Romano, G. et al. Mechanisms of regulation of cell adhesion and motility by insulin receptor substrate-1 in prostate cancer cells. Oncogene 20, 490–500 (2001) doi:10.1038/sj.onc.1204112
- prostate cancer
- cell adhesion
- cell motility
- signal transduction
Identification of novel mutations of Insulin Receptor Substrate 1 (IRS1) in tumor samples of non-small cell lung cancer (NSCLC): Implications for aberrant insulin signaling in development of cancer
Genetics and Molecular Biology (2019)
Reconstruction of global regulatory network from signaling to cellular functions using phosphoproteomic data
Genes to Cells (2018)
Calreticulin affects cell adhesiveness through differential phosphorylation of insulin receptor substrate-1
Cellular and Molecular Biology Letters (2014)
PLoS ONE (2014)
Molecular Profiling of Prostatic Acinar Morphogenesis Identifies PDCD4 and KLF6 as Tissue Architecture–Specific Prognostic Markers in Prostate Cancer
The American Journal of Pathology (2013)