Inhibition of SRC kinase activity does not restore sensitivity to anoikis in Ras-transformed intestinal epithelial cells

The intestinal epithelial cells (IECs) lining the lumen of the colon are normally attached to a complex matrix of proteins known as the Basement Membrane (BM). When this attachment to the BM is lost or prevented, normal IECs will undergo apoptosis (or “cell suicide”), in a process known as anoikis (in ancient Greek, “homelessness”) (1). On the other hand, colon cancer cells are not only able to survive in the absence of proper attachment to the BM but can also proliferate and invade surrounding and distant tissues, giving rise to metastases, i.e. malignant cells are by definition resistant to anoikis. Therefore, understanding the molecular mechanisms that regulate the execution of anoikis of IECs, and how these mechanisms are altered in colorectal tumor cells, may help us exploit this fundamental difference between normal and tumor cells and develop new therapies for the treatment of colon cancer. In fact, therapies based on restoring and/or enhancing the sensitivity of colorectal cancer cells to anoikis are expected to be fairly specific, since normally attached IECs would not be targeted by them, which would result in low toxicity.

One of the molecules whose function is most frequently altered (activated) in colon cancer patients is the non-receptor tyrosine kinase c-SRC (2). In the last few years, we have been studying some of the molecular mechanisms by which activated c-SRC can prevent anoikis in IECs. For that purpose, we use the previously characterized model cell line IEC18, which had been transformed with the oncoprotein v-SRC (a constitutively active mutant of SRC) (3,4). These v-SRC-transformed cells show levels of SRC kinase activity that are several times higher than those in the parental IEC18 cells (Fig.1). Importantly, the relative difference in the levels of c-SRC kinase activity found between colorectal cancer samples and surrounding normal stroma is very similar to that in our model (5). Even though our previous studies confirmed early observations that linked v-SRC to resistance to anoikis (1) and provided some cues on the molecular mechanisms involved in the process (4), they did not address whether physiological levels of c-SRC activity could also protect cells from anoikis. If resistance to anoikis were a “gain of function” phenotype related to carcinogenic hyperactivation of c-SRC, the inhibition of c-SRC activity as a pro-anoikic approach would only be applicable in cases that present such hyperactivation. Conversely, if normal c-SRC activity levels are also involved in regulating anoikis, then c-SRC could represent a more general target to enhance anoikis in detached cells. To start investigating this issue, we first analyzed the behavior of endogenous c-SRC activity in IEC18 cells in response to detachment. We found that after long periods in suspension culture (≥6 hours), the activity levels of c-SRC were significantly reduced (Fig 1). Thus, we were able to show there is a correlation between a decrease in SRC activity and anoikis, which led us to hypothesize that detached IECs might induce their own demise at least in part because they lose their c-SRC activity. One of the obvious implications of such hypothesis is that any intra- or extracellular process that inhibits the detachment-induced decrease in the activity of c-SRC would make the cells resist anoikis, even if it did not activate c-SRC activity above its basal levels in attached cells. Moreover, inhibition of c-SRC activity in these cases would, in principle, lead to an enhanced sensitivity to anoikis. We then decided to set out and explore this possibility in other well-established models of resistance to anoikis in intestinal epithelial cells.

Figure 1: Levels of c-SRC kinase activity in monolayer and suspension cultures of normal (IEC18) and three independently derived clones of v-SRC transformed intestinal epithelial cells. Top panel: The levels of c-SRC kinase activity in cells grown for 16 hours in monolayer (mon) or suspension (susp) were assayed based on the in vitro phosphorylation of an exogenous substrate of c-SRC (enolase). Levels of SRC autophosphorylation are also shown. Middle panel: A longer exposure of the upper panel is shown for improved appreciation of the downregulation of SRC activity in normal cells grown in suspension. Lower panel: Protein lysates from the same cells were probed by Western Blotting to determine the amounts of SRC in each case.

At the time, we were trying to understand the molecular mechanisms by which the Epidermal Growth Factor Receptor (EGFR) inhibits anoikis in IECs. The EGFR is a membrane associated receptor tyrosine kinase (RTK), which has been repeatedly associated with an anoikis resistant phenotype in diverse cellular systems (6-8). Briefly, binding of the EGFR to its ligand induces dimerization of the receptor, followed by a transautophosphorylation of each receptor molecule in specific tyrosine residues. These phosphorylated residues will now serve as docking sites for a large collection of signaling molecules that transduce the signal inside the cell, leading to a broad range of cellular effects, from rapid, transient changes in the activation state of diverse cellular components to the turning on and off of genes (9). Several ligands have been identified for the EGFR of which the Tumor Growth Factor (TGF)α was of most interest to us, given its role in contributing to the tumorigenic effect of activated Ras in IECs. The oncogenic activation of the small G protein Ras is another very common genetic alteration in colorectal cancers. In fact, early work in our laboratory focused on trying to uncover the molecular mechanisms by which the oncoprotein Ras confers IEC18 cells with their tumorigenic properties. One of our first findings was that Ras-transformed IEC18 cells show increased levels of expression and secretion of TGFα, and that the enforced reduction of TGFα production in these cells by antisense technology diminishes their ability to form colonies in soft-agar (i.e. to grow in anchorage-independent conditions) (3). Interestingly, one of the signaling molecules typically activated by the EGFR is Ras itself. What our results were then indicating is that a Ras-independent downstream effect of EGFR activation was required for full Ras-induced transformation.

Could such an effect be the activation of c-SRC? In fact, c-SRC is another well-known downstream effector of EGFR activation. As mentioned before, binding of the EGFR to its ligands leads to the phosphorylation of specific tyrosines in the receptor. One such phosphorylated tyrosine creates a binding site for a domain of c-SRC that would otherwise be blocking its active site. When this domain binds to the activated EGFR instead, it frees the active site of c-SRC, which is then turned on. Then, based on our previous findings, we decided to analyze whether the activation of c-SRC could be the aforementioned Ras-independent downstream effect of EGFR activation that led to anoikis resistance. Briefly, we were able to show that treatment of IEC-18 cells with TGFα partially restored the levels of c-SRC activity in suspended cells and protected them from anoikis (6). When the cells were treated with TGFα in the presence of PP1, a fairly specific Src family kinase inhibitor, the protective effect of TGFα was fully abrogated, which indicated that the partial restoration of c-SRC activity was responsible for the protective effect of TGFα.

Encouraged by these results, we sought to characterize the kinetics of c-SRC activity in Ras-transformed IEC18 cells. Knowing that Ras induces the expression and secretion of TGFα, and based on our results described above, we expected to see that after detachment from the BM, Ras-transformed IEC-18 cells would show higher levels of c-SRC activity than the non-transformed parental cells. However, much to our surprise, we could not detect any c-SRC kinase activity in Ras-transformed IEC18 cells, either in monolayer or suspension culture (Fig.2A), neither could we detect any expression of c-SRC by Western Blotting in these cells (Fig. 2A). Remarkably, other groups have since shown that other intestinal epithelial cell lines transformed by activated Ras show little to no c-SRC expression and activity (10).

Figure 2: A) Levels of c-SRC kinase activity in monolayer and suspension cultures of normal (IEC18) and Rastransformed intestinal epithelial cells. Upper panel: The levels of c-SRC kinase activity in cells grown for 16 hours in monolayer (mon) or suspension (susp) were assayed as in Figure 1. Middle panel: Protein lysates from the same cells were probed by Western Blotting to determine the amounts of c-SRC in each case. A saturated exposure is shown to stress the apparent lack of c-SRC in the Ras-transformed cells. Lower panel: The same protein blots as above were also probed for CDK4, which was used as a loading control. B) Effects of inhibiting c-SRC on anoikis of normal and Rastransformed intestinal epithelial cells. Levels of apoptosis under the indicated conditions were measured with a commercial kit for the detection of nucleosomal degradation (characteristic of apoptosis). These results are representative of two independent experiments with very similar results.

One “paradoxical” characteristic of the Ras-transformed cells is that they present almost undetectable levels of EGFR. The accepted explanation for such observation is that, when activated, the EGFR is rapidly internalized and turned over in lysosomes (11). Thus, the apparent lack of EGFR in the Ras-transformed cells simply reflects its rapid turnover, i.e. even though you cannot detect much EGFR in these cells, its downstream signals are constantly being delivered. We then entertained the possibility that something similar could be happening with c-SRC. If c-SRC was constantly being activated but rapidly degraded in the Ras-transformed cells, then it could have escaped our attempt to detect it by Western blotting or isolate it to measure its activity in vitro. In fact, a previous report had showed that a similar mechanism of c-SRC activity-induced degradation of c-SRC could be observed in certain cell types. Unfortunately, this observation is yet to be reproduced in other systems (7). We then decided to treat suspended Ras-transformed cells with the Src family inhibitor PP1. Rapidly turned-over c-SRC could have escaped our attempts to get it out of the cells, but should not escape the inhibitory action of PP1 being constantly present in the culture media. As shown in Figure 2B, we found that such PP1 treatment had no detectable effect on the survival of suspended Ras-transformed cells.

Our results support the conclusion that c-SRC activity is unnecessary for the TGFα-induced protection against anoikis in Ras-transformed cells. The simplest explanation for these results is that c-SRC activation is not part of the EGFR pathway that confers anoikis resistance in these cells. If activated Ras somehow inhibits the expression and/or activity of c-SRC, then the activation of the EGFR will necessarily induce resistance to anoikis in a c-SRC-independent manner. An alternative explanation is that, like in the parental cells, c-SRC is indeed being activated by the EGFR and is indeed contributing to the anoikis resistant phenotype, but through a pathway that would normally lead to Ras activation downstream of it. In such a model, c-SRC inhibition has no effect because the exogenous activated Ras feeds in the anoikis-inhibiting pathway downstream of the c-SRC inhibition.

There is yet a third possible explanation for our results. Even though several groups have shown that EGFR activation can protect diverse cell types from anoikis (7,8), we have not formally shown that EGFR function is necessary for the survival of the Ras-transformed cells in suspension. Our previous work had only demonstrated that Ras-transformed IEC-18 cells are dependent on TGFα overexpression/secretion to grow colonies in soft agar (i.e. anchorage-independent conditions) (3). Soft agar colony formation assays do not directly test the ability of cells to survive in suspension but rather their ability to proliferate in such conditions. Therefore, despite the abundant evidence supporting the role of the EGFR protecting diverse cell types against anoikis, we cannot rule out the following alternative scenario: even though SRC expression and activity levels were undetectable in our experiments with the Ras-transformed cells, these cells do have some SRC activity, and the fact that PP1 had no effect on the survival of the Ras-transformed cells is because EGFR in these cells is necessary for anchorage-independent growth but not for their survival. In other words, had we measured cell proliferation instead of survival in the suspended Ras-transformed cells, we may have found an inhibiting effect from the treatment with PP1.

In sum, these results suggest that transformation of IECs by Ras can either overcome the need for SRC activity in relaying the survival signals emanating from the activation of the EGFR or the need for such EGFR signals altogether. In either case, the main implication from these observations is that treatment of colorectal cancers that present activation of Ras with drugs targeting c-SRC activity may be futile as a pro-anoikic therapy. Whether or not the treatment of such tumors with drugs targeting the EGFR would induce anoikis and the regression of the disease, as discussed above, will depend on whether Ras transformation is only overcoming the need for c-SRC in EGFR signaling or if it is overcoming the need for EGFR signaling to prevent anoikis as a whole. Of course, further experimentation is needed to distinguish between these two possibilities.

Materials and Methods

Cell culture. As a model for normal IECs we used the IEC-18 cell line derived from rat small intestine. For models of Ras and Src activation, IEC-18 cells were stably transfected with activated HRas (V12N) and v-Src. Three independent clones were established from each transformation. Cells were grown in αMEM (5% FCS, 10μg/mL insulin, 0.36% m/v glucose, 1X GPS). For suspension cultures, regular tissue culture dishes were coated with an approximately 2mm thick layer of 1% sea plaque agarose in αMEM (1X GPS).

SRC in vitro kinase assay. An adapted standard protocol for an in vitro SRC kinase assay was used. Briefly, cells were lysed in modified RIPA buffer (50mM Tris-HCl pH 8; 120mM NaCl; 10mM EDTA; 200mM NaF; 1mM Na3VO4; 0.1% m/v SDS; 1% v/v NP40; 0.5% m/v DOC), and 1mg of total protein lysate per sample was incubated overnight with 1ug of anti-SRC antibody at 4°C (Upstate Biotechnologies – GD11), and then for 1 hour with 20μL of agarose-conjugated protein G beads (SIGMA) at 4°C. Beads were spun down (max speed-5sec) and washed three times in 200μL of lysis buffer and three more times in 200μL of 10mM Hepes (pH 8). Beads were then resuspended in 35μL of a master mix kinase reaction buffer (45mM Hepes pH 8; 150mM NaCl; 50mM MgCl2; 10μM Na3VO4; 2μM ATP and 10μCi of γ32P-ATP/sample) containing 0.04μg/uL acid-inactivated enolase per sample. To acid-inactivate enolase, an appropriate volume of a commercial suspension of rabbit muscle enolase (SIGMA), VOLeno, was mixed with 1 VOLeno of “solution 1” (60mM Hepes pH 8, 2.4mM DTT and 60% glycerol) and added to 2 VOLenos of “solution 2” (500mM Acetic acid). This mix was vortexed briefly and incubated at 37°C for 15 min. The reaction was then stopped by adding 4 VOLenos of “solution 3” (100mM Tris-HCl pH 8; 20mM MgCl2). The entire mix is added to the master mix of kinase reaction buffer described above.

Western Blotting. To analyze SRC expression, cell lysates were prepared as previously described. 30-35μg of protein lysate were run in a standard 8% SDS-PAGE gel. Blots were incubated overnight at 4°C with 2μg/mL anti-SRC antibody (GD11 – 1:500 dil). For loading controls, blots were incubated for 30 min at 50°C in stripping solution (62.5mM Tris-HCl pH 6.7, 2% SDS, 100mM 2-mercaptoethanol) and incubated for 1 hour at room temperature with anti-CDK4 50ng/mL (Santa Cruz).

Apoptosis assays. To measure levels of apoptosis, the Cell Death Elisa Kit (ROCHE) was used as per manufacturer’s instructions.


1. Frisch SM and Francis H (1994). J Cell Bio 124(4):619-26.

2. Irby R et al. (1997). Oncogene 19(49):5636-42.

3. Filmus J et al. (1995). Oncogene 8(4):1017-22.

4. Coll ML et al. (2002). Oncogene 21(18):2908-13.

5. Talamonti MS et al. (1993). J Clin Inv 91:53-60.

6. Rosen K et al. (2001). J Biol Chem 276(40):37273-79.

7. Jost M et al. (2001). Mol Biol Cell 12(5):1519-1527.

8. Schulze A et al. (2001) Genes Dev 15(8):981-994.

9. Kari C et al. (2003). Can Res 63(1):1-5.

10. Dehm S et al. (2001). FEBS Lett 487(3):367-71.

11. Burke P et al. (2001). Mol Biol Cell 12(6):1897-1910.

12. Hakak Y and Martin GS (1999). Curr Biol 9(18):1039-42.

13. Jost M et al. (2001). Mol Biol Cell 12(5):1519-27.

14. Schulze A et al. (2001). Genes Dev 15(8):981-94.

Share your thoughts

Leave a Reply

You must be logged in to post a comment.