Colorectal Tumors Are Effectively Eradicated by Combined Inhibition of β-Catenin, KRAS, and the Oncogenic Transcription Factor ITF2

The targeting of oncogenic proteins driving human cancers represents a fascinating strategy to overcome the obvious limitations of cytotoxic drugs. The development of imatinib as a specific inhibitor of Bcr/Abl represents the most successful example where specific oncogene targeting dramatically changed the clinical course of a human cancer (1). This case summarizes several factors needed for the success of such an attempt: the causal relationship between target and disease; early involvement of the target in the process of malignant transformation; and the activity and specificity of the inhibitor. However, in tumors where the molecular pathogenesis is brought out by several oncogenes resulting in the aberrant activation of specific pathways, an important therapeutic effect may need the combined inhibition of more than a single target.

Wnt signaling is implicated in a variety of cellular processes, including proliferation, differentiation, survival, apoptosis, and cell motility. Its constitutive deregulation results in the development of various tumors (2). The canonical Wnt pathway regulates the stability of the proto-oncogene β-catenin (β-cat), and its activation leads to β-cat/T-cell factor (TCF)–dependent transcription (3). The level of cytoplasmic β-cat is tightly controlled by a degradation complex containing the tumor suppressor adenomatous polyposis coli (APC). In nonstimulated cells, this complex mediates the degradation of cytoplasmic β-cat through a multistep process (4). On Wnt signal activation, the destruction complex is inhibited and β-cat translocates into the nucleus, where it interacts with TCF transcription factors. TCF-4 (TCF7L2) is the most highly expressed TCF factor in colorectal carcinoma (CRC) cells and is involved in the maintenance of undifferentiated intestinal crypt epithelial cells (5). Dominant-negative TCF-4 induces cell cycle arrest and differentiation (6) and restores the epithelial cell polarity of colorectal cancer cells, converting the cell population into a single layer of columnar epithelium (7). This indicates that CRC cells still require accumulation of β-cat for maintenance of proliferation, depolarization, and dedifferentiation. A decrease of β-cat levels in established colon cancer cell lines reduces proliferation and induces differentiation but does not lead to apoptosis (8–10).

The basic helix-loop-helix (bHLH) transcription factors play a critical role in cell fate determination and differentiation in a variety of tissues (11). bHLH proteins can be arranged into two distinct classes depending on their dimerization specificity and expression pattern. Class I bHLH proteins are ubiquitously expressed and cooperate with tissue-specific class II bHLH factors to regulate gene expression. Class I proteins consist of ITF2 (12, 13), HEB, and the differentially spliced products of the E2A gene (E12 and E47). ITF2 encodes for two transcripts, ITF2A and ITF2B, which have activating and repressing transcriptional activities, respectively (14). In particular, ITF2B has been shown to inhibit the transcription of various genes related to cell differentiation by sequestering tissue-specific bHLH proteins (14–16). Previous studies showed that ITF2B is a downstream target of the Wnt/β-cat pathway and is consistently activated in β-cat–transformed cells (17). ITF2B is strongly overexpressed in primary ovarian carcinomas with defective β-cat regulation. Moreover, ITF2B levels were significantly reduced on wild-type APC reexpression in CRC cells. Finally, aberrant expression of ITF2B in epithelial cells promoted neoplastic transformation (17). However, Herbst and colleagues (18) surprisingly found that ITF2B is strongly induced in colorectal adenomas but is downregulated in about 60% to 70% of carcinomas, often as a result of 18q loss of heterozygosity. Thus, the question remains open as to whether ITF2B is an oncogene or a tumor suppressor.

KRAS-activating mutations are found in ∼50% of colon cancers. In the majority of cases, KRAS mutations arise very early in tumor history and are associated with mutations that activate the Wnt/β-cat pathway (19). A few transgenic mouse models have been described highlighting the importance of both Wnt/β-cat and KRAS signaling in colon carcinogenesis. In these studies, the Wnt/β-cat and KRAS pathways were shown to cooperate in inducing CRC initiation and progression (20–22). Moreover, APC has been shown to block the Ras pathway in CRC cells by interfering with GTP loading of KRAS (23). Several genes have been shown to be regulated by both the Wnt/β-cat and KRAS pathways, including BIRC5/Survivin, COX2, MYC, CCND1/cyclin D1, and MYB.

Thus far, independent targeting of single oncogenic events in CRC has met limited therapeutic success (6, 9, 24, 25). In an effort to better understand the contribution of Wnt/β-cat and KRAS pathway defects in colon cancer, a simultaneous inhibition of multiple oncogenic events was performed. We studied the effects of inhibiting the two main pathways involved in malignant transformation of human cells (Wnt/β-cat and KRAS) and analyzed the role of ITF2 in CRC cell growth and survival. We used doxycycline (Dox)-inducible short hairpin RNA (shRNA) to downregulate β-cat, ITF2, and KRAS proteins in CRC cells and report here the in vitro and in vivo results obtained, as well as a signal transduction pathway analysis, which led to the identification of genes commonly regulated in the different shRNA combinations.

Ls174T and HCT-116 cell lines were purchased from the American Type Culture Collection, where they are routinely verified using polyphasic (genotypic and phenotypic) testing to confirm identity. Ls174T and HCT-116 cells carry mutations in the CTNNB1 gene that stabilize β-cat protein and mutations in the KRAS gene that cause a constitutively GTP-loaded, active state of KRAS protein. Ls174T-TR cells stably expressing the Tet-repressor were obtained from Dr. Hans Clevers (The Hubrecht Institute, Utrecht, the Netherlands). To generate the corresponding HCT-116-TR cell line, HCT-116 cells were transfected with pcDNA6/TR plasmid and selected with blasticidin. Ls174T-TR and HCT-116-TR cells were used for subsequent transfection of the inducible shRNA constructs, as described in Supplementary Materials and Methods.

In vivo studies were approved by the McGill Animal Care Committee and by the Ethics Committee for Animal Experimentation of the National Cancer Institute of Milan. Five-week-old female nu/nu mice were obtained from Charles River and housed in clean pathogen-free rooms in groups of five in cages containing microisolator tops. Five million cells in 0.2 mL of PBS were injected s.c. in the right flank of the mice. Tumor volumes were measured every 3 to 4 days using a caliper and the following equation: tumor volume = d² × D/2, where d and D are the shortest and longest diameters of the tumor, respectively. Mice were monitored for signs of disease (weight loss, adenopathies). Survival was determined from the day of tumor injection to the day of euthanasia. Therapeutic efficiency of the shRNA was assessed by addition of Dox in the drinking water (0.2 mg/mL Dox, 5% sucrose), which was changed twice a week. To analyze in vivo β-cat, ITF2, and KRAS protein levels, tumor-bearing mice were killed and nonnecrotic tumor tissue was extracted and homogenized in a 5-fold volume of loading buffer [50 mmol/L Tris-HCl (pH 6.8), 2% SDS, and 5% β-mercaptoethanol]. After sonication, samples were centrifuged at 15,000 × g for 15 minutes, heated at 95°C for 10 minutes, and analyzed by Western blot.

Apoptosis pathway–focused gene expression profile of Ls174T cells, before and after Dox treatment, was performed with the RT² Profiler PCR Array System (SABiosciences) using the Human Apoptosis array according to the manufacturer's recommendations. Cells were treated with vehicle or Dox for the time periods indicated in the figure legends and then harvested. Total RNA and proteins were extracted with TRIzol reagent (Invitrogen). Equal amounts of total proteins were run in Western blot to confirm downregulation of target proteins in Dox-treated samples. RNA was further purified using the RT² qPCR-Grade Total RNA Isolation Kit (SABiosciences) and cDNA was synthesized with the RT² First Strand Kit. Real-time PCR was performed on an Mx3005p machine (Stratagene) using recommended conditions. Raw data were analyzed using the PCR Array Data Analysis Software (26). All expression values were normalized to the average of five housekeeping genes. Fold-change values, defined as the ratio between Dox sample and control sample, were reported for each cell line in the Heat Map generated with HeatMap Explorer software.

Cells were cultured in 96-cell well plates in replicates of six and treated for 5 days with 1 μg/mL Dox or vehicle. At each time point, 1 μCi of [³H]thymidine (Amersham) was added to each well 8 hours before harvesting onto glass fiber filters by a Tomtec automated cell harvester. Incorporation of [³H]thymidine was measured using a filter scintillation counter (1430 MicroBeta).

For apoptosis determination, the cells were harvested, washed with PBS, and resuspended in Annexin binding buffer. Approximately 10⁵ cells were then stained with Annexin V and propidium iodide and analyzed by flow cytometry as described (27). An aliquot of cells was kept for caspase-3 activity measurement with the Caspase-Glo assay kit (Promega) as described (27).

The cells were harvested, washed in PBS at 4°C, resuspended in lysis buffer [50 mmol/L Tris-HCl (pH 7.4), 1% Triton X-100, 5 mmol/L EDTA, 150 mmol/L NaCl, 1 mmol/L Na₃VO₄, 1 mmol/L NaF, 1 mmol/L phenylmethylsulfonyl fluoride, and protease inhibitor cocktail (10 μmol/L benzamidine-HCl and 10 μg/mL each of aprotinin, leupeptin, and pepstatin A)] and incubated on ice for 30 minutes. Lysates were then clarified by centrifugation at 13,000 × g for 15 minutes at 4°C. The protein concentration of cell lysates was determined using the bicinchoninic acid protein assay (Pierce). Equal amounts (30 μg) of total proteins were loaded on SDS-PAGE, transferred onto a nitrocellulose membrane, and probed with primary antibodies (see Supplementary Materials and Methods) overnight at 4°C. Horseradish peroxidase–conjugated secondary antibodies were incubated for 1 hour at room temperature. Proteins were visualized by chemiluminescence as recommended by the manufacturer (Amersham).

SigmaPlot and GraphPad-Prism programs were used to analyze the data and plot curves. Two-tailed unpaired t test was used to determine statistical significance of the differences between data sets, where appropriate. Kaplan-Meier analysis was used to plot tumor-free survival curves.

We used published Dox-inducible shRNA to confirm previous findings (8) showing that specific and substantial downregulation of β-cat protein levels (up to 90%, as evaluated by Western blotting) can be achieved within 4 days of culture in Ls174T cells (Fig. 1B). Silencing of β-cat caused a concomitant downregulation of the bHLH transcription factor ITF2, as expected from previous data showing that ITF2 is a β-cat/TCF target (17). ITF2 was expressed in all tested colon cancer cell lines, as assessed by real-time PCR and Western blot (Supplementary Fig. S1A and B). To address whether ITF2 expression may have a role in regulating the proliferation of colon cancer cells, shRNA was used to decrease ITF2 expression in Ls174T cells. Within 3 days of Dox treatment, ITF2 protein levels were greatly reduced and reached maximal inhibition (95%) by day 5 (Fig. 1C). ITF2 expression after 5 days of shRNA induction was significantly lower compared with cells that carry the β-cat shRNA alone (compare Fig. 1B and C). Therefore, stronger downregulation of ITF2 expression was achieved by direct anti-ITF2 shRNA. The mRNA level of ITF2 was downregulated after 2 days of treatment, as evaluated by real-time PCR (Supplementary Fig. S2). The same inducible-shRNA strategy was used to knock down KRAS in Ls174T cells. KRAS, but not actin, protein levels were strongly reduced (up to 80%) after treatment with Dox (Fig. 1D).

We investigated the effects of β-cat, ITF2, or KRAS downregulation on cell growth. Maximal proliferation inhibition (88% for β-cat, 73% for ITF2, and 45% for KRAS) was reached in all cases at day 3 to 4 in Ls174T cells (Fig. 1B–D). Control cells expressing empty vector (Fig. 1A) or scrambled RNAi (data not shown) did not show perturbation of their growth rate on Dox treatment. These data show that inhibition of either one of these three genes has a significant effect on CRC cell proliferation.

We then assessed the effects of the combined targeting of two genes. Double-shRNA Ls174T clones that simultaneously silence β-cat/ITF2 or β-cat/KRAS after exposure to Dox were established. In these double-shRNA cells, decreased expression of both targeted proteins was observed after 48 to 72 hours of Dox treatment (Fig. 1E and F). In vitro cell proliferation assays showed that combined β-cat/ITF2 silencing inhibited cell growth by approximately 99% after 5 days of treatment with Dox (Fig. 1E). In double-shRNA β-cat/KRAS cells, proliferation was inhibited by approximately 97% after 5 days of treatment, compared with untreated control (Fig. 1F). Similarly, Dox-treated double-shRNA clones were 100% inhibited in anchorage-independent growth assay in soft-agar medium (data not shown). To confirm these findings in an independent CRC cell line, we reproduced the data in HCT-116 cells (Supplementary Fig. S3). As in LS174T cells, double β-cat/KRAS and β-cat/ITF2 silencing caused significantly greater cell growth inhibition compared with single targeting in HCT-116 cells, as determined by tritiated thymidine uptake (Supplementary Fig. S3C) or by soft-agar growth and MTS assays (data not shown).

Taken together, these results indicate that concurrent downregulation of two oncogenic factors can increase proliferation inhibition in CRC cells.

Based on these results, it was important to address the issue of whether the decrease in [³H]thymidine incorporation seen in the shRNA-transfected cells was followed by induction of apoptosis. Figure 2A shows the results of a representative experiment in which the combined downregulation of β-cat and KRAS or β-cat and ITF2 induced early apoptosis in approximately 45% of treated cells, compared with less than 5% in controls or in cells in which β-cat, ITF2, or KRAS was singly downregulated. Figure 2B shows the overall results from three experiments: the independent targeting of either β-cat, ITF2, or KRAS did not lead to increased cell death over controls. In contrast, apoptosis was readily observed in β-cat/ITF2 (42 ± 3%) or β-cat/KRAS (45 ± 6%) double-shRNA–transfected Ls174T cells. The apoptotic rate observed in these two groups was substantially higher than that observed when a single shRNA was induced (P < 0.01), even when considering the sum of the values obtained with the two single shRNAs. These results were further confirmed using induction of caspase-3 activity (Fig. 2C) and sub-G₁ analysis (data not shown). Similar results were obtained in HCT-116 cells, in which combined shRNA showed superior lethality compared with single silencing (Supplementary Fig. S3D). These data strongly point to a synergistic activity of two inhibited genes in the development of apoptosis in treated cells.

To determine the potential effects of shRNAs on colon cancer cell growth in vivo, equal numbers of double-shRNA β-cat/ITF2 cells, β-cat/KRAS cells, cells transfected with single β-cat, ITF2, or KRAS shRNA, or cells transfected with empty vector (control) were injected s.c. into nude mice. When tumors became palpable, half of the mice from each group were randomized to receive continuous administration of Dox in the drinking water. The growth of tumors was measured every 3 days. Control (Fig. 3A) or single shRNA cells (Fig. 3B–D) did not show significant growth inhibition effects after Dox treatment. In contrast, both β-cat/ITF2 and β-cat/KRAS shRNA cell lines showed a significant reduction of tumor growth under Dox treatment, compared with untreated mice (Fig. 3E and F), with treated tumors showing almost no growth over >40 days of treatment. In vivo modulation of targeted genes was verified in each case, by real-time PCR and Western blot, in excised tumors (Supplementary Fig. S4).

The effects of gene silencing on large tumors were studied in nude mice carrying bulky (>1 g) β-cat/ITF2 double-shRNA–expressing tumors. As shown in Fig. 4D, almost complete regression (>95%) of tumors occurred in Dox-treated tumors, with no significant regrowth up to 200 days of observation. By contrast, single shRNA and control cells did not modify their growth rate on Dox treatment (Fig. 4A–C). Less dramatic effects, consisting in growth delay rather than regression, were observed with β-cat/KRAS double-shRNA cells, when Dox was given 22 days after cell injection (Fig. 4E).

Kaplan-Meier curves were obtained by scoring the time of appearance of a measurable tumor mass in mice treated 24 hours after tumor cell injection (Fig. 5). The results indicate that Dox administration in both double-shRNA cell populations (β-cat/ITF2 and β-cat/KRAS) protected mice against tumor growth (Fig. 5E and F), with a 40% tumor-free survival rate in both groups, whereas mice injected with single shRNA cells or control cells developed tumors in 90% to 100% of cases with no differences between treated and untreated mice.

A similar trend was obtained using HCT-116 cells: double β-cat/KRAS and β-cat/ITF2 RNAi xenografts grew significantly more slowly in nude mice under Dox compared with any single shRNA clone (Supplementary Fig. S3).

Taken together, these data indicate that targeting two proteins at the same time may exert a strong antitumor effect in vivo. They also suggest that the development of apoptosis in vitro is associated with tumor responses in vivo.

To study downstream gene expression changes consequent to shRNA induction, Western blot analysis of Dox-treated and untreated cells was performed (Fig. 6A; Supplementary Fig. S3A). As expected, Dox treatment did not affect protein expression in cells expressing the control construct. In agreement with the data found in the literature, c-myc expression was downregulated by all shRNA cell lines. Survivin expression decreased in double-shRNA cells and in ITF2 single shRNA, whereas Sp1, which is known to induce survivin transcription (28), was downregulated in all the double-shRNA cells as well as in β-cat shRNA cells. c-Src and p27^KIP1 showed opposite modulation in all cells that silenced β-cat. Surprisingly, KRAS shRNA did not block extracellular signal–regulated kinase (ERK) phosphorylation in Ls174T or HCT-116 cells. Persistent activation of the mitogen-activated protein kinase (MAPK) pathway was confirmed by phospho-MAPK/ERK kinase (MEK)-1/2 (data not shown). In line with these data, treatment of Dox-induced β-cat shRNA cells with a BRAF (29) or MEK1/2 (30) inhibitor did not cause significant apoptosis (data not shown), suggesting that the RAF/MEK/ERK pathway is not a major transducer of KRAS oncogenic signal in these cells. Therefore, we evaluated the possible involvement of alternative KRAS-dependent pathways, such as phosphatidylinositol 3-kinase/AKT and RalA/phospholipase D. Whereas phospho-AKT (Ser473) levels did not change, Fos (a known RalA transcriptional target; refs. 31–33) expression decreased on KRAS silencing (Fig. 6A), suggesting that RalA may be the main effector of KRAS in CRC cells.

Interestingly, the mTOR effector molecule p70S6K was downregulated only in β-cat/ITF2 and β-cat/KRAS double-shRNA cells on Dox treatment. This result suggests that a common pathway may be involved in CRC cell death following double-gene targeting.

Surprisingly, in both Ls174T and HCT-116 cells, KRAS was strongly downregulated in the double β-cat/ITF2 shRNA transfectant, making these cells, as a matter of fact, a “triple knockdown.” This may explain the superior in vivo inhibitory effect observed for this combination of shRNA on large tumors. In contrast, β-cat and β-cat/KRAS shRNA cells only partially decreased ITF2 levels.

To gain insight into the regulation of pathways leading to apoptosis in Dox-treated double-shRNA cells, we ran a quantitative PCR gene expression profiling of 84 apoptosis-related genes (Fig. 6B). Quantitative comparisons are presented in the Supplementary Table. As shown in the heat map of Fig. 6B, cells that silenced ITF2, either alone or in combination with β-cat, showed a markedly different gene regulation pattern compared with the other cell lines, showing many genes to be significantly upregulated. However, single ITF2 downregulation was insufficient to induce apoptosis and to cause tumor regression in vivo. These data suggest that β-cat/ITF2 silencing and β-cat/KRAS silencing induce cell death through different apoptotic pathways. Despite the overall difference, some response shared by β-cat/ITF2 and β-cat/KRAS shRNA could be found: both cells were able to strongly downregulate the BCL2 gene and to moderately upregulate FAS. In addition, β-cat/KRAS shRNA cells significantly decreased BCL2L1 transcript (encoding for the antiapoptotic Bcl-xL protein), whereas β-cat/ITF2 cells uniquely upregulated proapoptotic Bim (BCL2L11). These results suggest that the balance of proapoptotic and antiapoptotic genes was shifted toward apoptosis in both double-silenced cells, although through different pathways.

Cancer represents the result of a multistep process. Even neoplasias, considered as “simple models,” are now known to require multiple genetic lesions to display a full-blown malignant phenotype. Recent experience with kinase inhibitors indicates that the different therapeutic value of target genes resides in their position along the transformation/progression process. Early pathogenic events are more likely to provide relevant therapeutic targets. The elucidation of the different genetic events leading to human cancer is only partial. High-throughput sequencing and other techniques hold the promise for rapidly increasing our knowledge in the field. Recent studies provide us with a fairly stable number of genes becoming altered in malignant cells: 12 in the case of acute myelogenous leukemia, 14 for breast cancer, and 15 for CRC patients (34). Although the therapeutic potential of a target depends on several factors such as its activity (transforming versus tumor suppressor), position in the transformation hierarchy (early versus late occurring), and availability of specific inhibitors, the identification of genetic alterations in a given cancer type (or patient) will supply plenty of potential therapeutic targets.

When single inhibition of the three genes analyzed in this report (β-cat, ITF2, or KRAS) was performed in CRC cells, strong proliferation inhibition was obtained, but no substantial induction of apoptosis or any in vivo growth inhibition was noted. When the combined inhibition of two genes (either β-cat/ITF2 or β-cat/KRAS) was induced in vitro or in vivo, important biological consequences ensued. Proliferation inhibition was more complete than with single gene silencing, but more importantly, massive apoptosis was noted, which cannot be explained by the mere addition of the apoptosis values induced with the single inhibition of the two involved genes. Interestingly, apoptosis induction was also associated with tumor responses in vivo, which include growth inhibition, regression of large tumors, and protection of animals from the growth of injected tumor cells. Thus, the induction of apoptosis in vitro seems to predict for the subsequent development of in vivo responses.

It is somewhat surprising that the MEK/ERK pathway was not inactivated on KRAS silencing in our cells. The fact that we observed the same result in two different cell lines rules out any clonal effect. Consistent with our data, previous studies found that ERK was not hyperactivated in endogenous KRAS^G12D-driven intestinal tumors (20, 35). Similarly, Scholl and colleagues (36) reported that the phosphorylation status of MEK and ERK does not correlate with KRAS dependency. Therefore, additional KRAS-driven pathways, such as RalA, might be responsible for tumorigenesis, as suggested both by our results and by previous findings (37).

The analysis of signal transduction in the various transfected cells allows one to draw some preliminary conclusions on the pathways that could be critical in generating the responses observed in our study. Initial analysis pointed to p70S6K (an effector of the mTORC1 pathway) as a protein specifically downregulated in the two double-shRNA transfectants. Modulation of p70S6K activity has recently been described as a major effector of survival signals in mutant KRAS–dependent cancer cells (36), and p70S6K expression is greatly increased in KRAS^D12/APC^Min mice (38). Interestingly, the antitumor effect of farnesyltransferase inhibition has been linked to inactivation of p70S6K (39). Further study in apoptosis-involved genes identified the downregulation of BCL2 and the upregulation of FAS as being present only in both double-shRNA cells. Thus, it seems that the hallmark of the in vivo response consists in the ability of shifting the proapoptosis/antiapoptosis balance toward the first one. This effect can be achieved by modifying the levels of the same molecules (p70S6K) and by other modifications developing only in one double-shRNA transfectant, such as Bim upregulation in β-cat/ITF2 cells and Bcl-xL downregulation in β-cat/KRAS transfectants. Whereas our results suggest these genes to be relevant for the therapeutic effects observed, the biological effects of their specific downregulation were not experimentally validated, as the observed phenotype is probably caused by several rather than a single modification in the expression levels of the described genes.

The overall comparison of expression profiles for anti–β-cat, anti-ITF2, anti-KRAS shRNA, and related combinations shows important differences among the various transfectants, with a prevalence of upregulated genes in anti-ITF2 transfectants (in keeping with the described transcriptional repressor activity of ITF2B; ref. 17) and of downregulated ones in anti-β-cat and anti-KRAS cells. Nonetheless, it seems that the net balance is shifted decisively in favor of apoptosis in the double transfectants, although through different pathways; for example, anti-ITF2 shRNA cells showed a very potent upregulation of a prosurvival gene (BIRC8), which, however, is repressed in the double-shRNA transfectant. Such a pattern is not present in KRAS versus β-cat/KRAS shRNA transfectants, in which BIRC8 is constantly downregulated.

These experiments also produced an unexpected result: Combined β-cat/ITF2 shRNA caused the downregulation of KRAS, thus effectively acting as a triple knockdown. To our knowledge, this represents the first description of KRAS downregulation operated by the simultaneous inhibition of β-cat and ITF2. These results also epitomize how different the effects can be when a double inhibitory strategy is applied: Neither β-cat nor ITF2 single shRNA achieved KRAS inhibition. Although ITF2 was confirmed to be a Wnt/β-cat target, complete suppression of its expression could only be attained by anti-ITF2 shRNA. Hence, the combined β-cat/ITF2 silencing should, in theory, achieve a better inhibition of the oncogenic pathway than β-cat shRNA alone because the downstream ITF2 molecule is completely blocked. KRAS expression was not known to be controlled by either gene. The mechanisms by which this phenomenon is determined as well as the other biological activities of ITF2 inhibition (such as the downregulation of phospho-ERK) will require further studies.

The clinical application of our data is not immediate, mainly for two reasons: First of all, pharmacologic inhibitors of these targets are not clinically available yet. It is, however, encouraging that inhibitors for oncogenic proteins not belonging to the kinase family have recently been developed (40), raising hopes that such molecules can be designed. Second, the biological consequence of multiple inhibitions of β-cat, ITF2, and KRAS in normal tissue remains to be evaluated. However, according to the “oncogene addiction” model, tumor cells should be more sensitive to selected survival pathway inhibition compared with normal cells (41). In view of the results published by Herbst and colleagues (18), showing downregulation of ITF2 at the adenoma-to-carcinoma transition in a significant subset of colon cancers, it may be questioned whether ITF2 represents a valuable target in CRC. Our study strongly suggests that at least in those tumors where its expression is maintained, ITF2 inhibition has dramatic effects on cell growth and survival when combined with β-cat silencing. Therefore, our data suggest that ITF2 relays an oncogenic signal, as reported by Kolligs and colleagues (17). Possibly, tumors that lose ITF2 expression during transition to carcinoma may have acquired additional abnormalities. This scenario once again highlights the importance of personalized molecular diagnosis for patient management. Furthermore, our results are in contrast to the expression data reported by Herbst and colleagues in cell lines (18). We do not know the reason for this discrepancy. We observed ITF2 expression in six CRC cell lines using two different techniques, although at variable levels. Moreover, ITF2 expression has been previously reported in HT-29 and DLD-1 cells (17).

The data presented here have potential clinical implications: They show in fact that the targeting of a single gene (β-cat or KRAS) or pathway may be insufficient to obtain substantial therapeutic effects in CRC, even if that gene or pathway is causally involved in the generation or maintenance of the malignant phenotype. Simultaneous targeting of different genes or pathways involved in malignant transformation could instead produce a decisive improvement in the therapeutic results of these approaches.

No potential conflicts of interest were disclosed.

We thank Dr. Hans Clevers (The Hubrecht Institute, Utrecht, the Netherlands) for providing Ls174T cells expressing TetR, Ls174T cells expressing β-catenin shRNA, and the pTER vector.

Grant Support: Italian Association for Cancer Research (IG-4637) and the Italian Government MIUR-COFIN and PRIN programs; CIHR, NCI-C, CFI.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.