Gene Ther Mol Biol Vol 8, 385-394, 2004

*Correspondence: A. Ivana Scovassi, Istituto di Genetica Molecolare CNR, Via Abbiategrasso 207, 27100 Pavia, Italy; Tel +39-0382-546334; Fax +39-0382-422286; E-mail: scovassi@igm.cnr.it

Key words: Antisense, apoptosis, cancer, c-myc, phosphorylation, TFO

Abbreviations: antisense oligonucleotides, (AS-ODN); disialoganglioside, (GD₂); Oligonucleotides, (ODNs); Ribonucleoprotein, (RNP); triple helix-forming oligonucleotides, (TFOs)

It is generally assumed that the efficacy of anticancer drugs may be related to cell proliferation control and/or to the activation of the apoptotic pathway(s). Among the mediators of such processes, the c-myc proto-oncogene controls the balance between proliferation and death, thus playing a crucial role in different cell pathways leading to opposite effects (Prendergast, 1999; Amati et al, 2001; Eisenman, 2001; Nasi et al, 2001; Pelengaris et al, 2002; Pelengaris and Khan, 2003). In this respect, c-myc could be represented as Janus, the old Roman deity with two faces who presides over everything by regulating cell proliferation and cell death (Figure 1).

A simplified view of the activities of c-myc is shown in Figure 2. In normal cells, c-myc expression is tightly controlled by mitogenic stimuli and appears to be necessary, and in some instances sufficient, to induce cells to enter the S phase of cell cycle and to proliferate, and to respond to differentiative stimuli (Hoffman and Liebermann, 1994). Translocation and amplification of the c-myc gene as well as increased half-life and overexpression of the oncoprotein, which have been observed in many tumors, promote tumorigenesis (Spencer and Groudine, 1991; Marcu et al, 1992).

Deregulation of c-myc occurring in a broad range of human cancers is often associated with poor prognosis (Pelengaris et al, 2002). The molecular mechanisms for the frequently observed deregulation of c-myc in human cancers could depend on the fact that c-myc overexpression may antagonize the pro-apoptotic function of p53 (Ceballos et al, 2000). c-myc controls or affects other processes relevant to tumorigenesis, e.g. it can promote transformation by its ability to induce the expression of telomerase, thus bypassing telomere erosion and facilitating immortalization (Drissi et al, 2001).

Different factors may regulate in distinct ways c-myc-promoted cell transformation (O'Hagan et al, 2000). Among them, Bim acts as a suppressor of Myc-induced lymphomagenesis (Egle et al, 2004); non-peptide antagonists of Myc/Max dimerization inhibit c-myc-induced transformation (Berg et al, 2002); the ATM-related domain of TRRAP protein, which is involved in transcriptional regulation and chromatin structure, modulates c-myc-dependent oncogenesis (Park et al, 2001).

c-Myc protein is a member of the helix-loop-helix leucine zipper family of transcription factors that bind to a DNA motif called "E-box", which consists of the consensus sequence CACGTG. Efficient binding of c-Myc to an E-box requires the heterodimerization with its partner Max, another member of this family. Myc function is antagonized by the Mad protein, which can also dimerize with Max and bind to E-boxes (Amati et al, 2001; Baudino and Cleveland, 2001; Zhou and Hurlin, 2001). Since the main activities of Myc strictly depend on its dimerization with Max, the inhibition of such interaction may affect different processes. Indeed, small molecules acting as inhibitors of Myc/Max dimerization were effective in counteracting the oncogenic activity of Myc (Berg et al, 2002).

c-myc initiates a transcriptional program that controls hundred of genes belonging to different functional categories of myc targets. Some of them can be considered as direct targets, others are indirectly regulated. The investigation of the nature of the interaction among c-Myc network members revealed that it could be modulated through the formation of distinct sub-nuclear structures localized in specific compartments (Yin et al, 2001).

Figure 1. Representation of the oncogene c-myc as the double-headed Janus deity. Looking in the direction of both cell proliferation and death, c-myc controls the basic life processes.

Figure 2. Regulation of different processes by c-myc in normal cells. Effect of c-myc deregulation in promoting cancer.

To date, the search for c-myc targets did not provide conclusive data. A still growing list of proteins regulated by c-Myc is reported and discussed in many reviews (Dang, 1999; Sakamuro and Prendergast, 1999; O'Hagan et al, 2000; Eisenman 2001; Levens, 2002, 2003; Fernandez et al, 2003; Nilsson and Cleveland, 2003, 2004; Patel et al, 2004). A variety of molecular, biological and genetic approaches were devised to identify the mRNAs induced or repressed by c-myc. Recent advances in proteomics and microarray technology allowed genome-wide studies of mRNA transcripts responsive to c-Myc (Schuhmacher et al, 2001; Shiio et al, 2002; Watson et al, 2002; Fernandez et al, 2003; Orian et al, 2003).

The observation that c-myc null fibroblasts are resistant to apoptosis highlighted the essential pro-apoptotic role of this oncogene (Chang et al, 2000). It is generally assumed that c-myc promotes apoptosis by sensitizing cells to a variety of insults rather than by acting as a direct death effector. Yu et al (2002) carried out a genome-wide survey for myc-mediated gene expression under apoptotic conditions. Isogenic Rat-1 cell lines that either overexpress or lack c-myc, were treated with etoposide, which induced apoptosis at an extent that depend upon the level of c-myc. The analysis provided the identification of a cluster of genes that respond to etoposide and are highly dependent on the cellular myc status. Moreover, the results revealed also that the existence of c-myc-independent genes involved in the apoptotic pathway.

Although a detailed understanding of the signalling pathways by which c-myc elicits apoptosis is still lacking, different factors have been shown to modulate c-myc-induced apoptosis. As first shown by Fanidi et al (1992) and Bissonnette et al (1992), the ability of c-myc to promote apoptosis can be suppressed by the overexpression of bcl-2; the same effect was obtained by the suppression of the pro-apoptotic factor Bax (Mitchell et al, 2000). Ionizing radiation-induced apoptosis can be increased by the activity of c-Myc in suppressing Bcl_XL, thus suggesting a strategy in desensitizing tumor cells to DNA damage-induced apoptosis (Maclean et al, 2003). The transcriptional repressor Mad1, which regulates negatively cell proliferation, has an inhibitory effect on c-myc-mediated apoptosis and proliferation (Gehring et al, 2000). Using RNA stable interference (siRNA), Nilsson and Cleveland (2004) showed that Mnt, a myc antagonist (Hurlin et al, 2004), triggers apoptosis via the myc target ODC. A similar indirect effect was described for the complex formed by the c-myc-negative regulator MBP-1 (c-myc promoter-binding protein 1), and MIP-2A (MBP-1-interacting protein), which in turn regulates negatively the MBP-1 activity and the induction of apoptosis (Ghosh et al, 2001).

A synergy between c-myc and different death receptors, leading to the release of cytochrome c from mitochondria, was shown (Klefstrom et al, 2002). Remarkably, it has been reported that the gene for cytochrome c, which is required for apoptosis, is a direct target of c-myc and that c-Myc binds to it (Morrish et al, 2003). The analysis of the apoptosis induced in melanoma cells after c-myc down-regulation revealed that this process occurs through the specific depletion of the levels of glutathione (Biroccio et al, 2002).

In contrast to the pro-apoptotic function usually ascribed to c-myc, it has been shown that c-myc could contribute to block apoptosis under some conditions. In lymphoid CEM cells, treatment with oxysterols reduces c-Myc protein expression level before promoting apoptosis (Ayala-Torres et al, 1999), thus suggesting that the negative regulation of c-Myc does not inhibit the activation of apoptosis by steroid compounds.

c-Myc is a highly unstable phosphoprotein with a half-life of about 15-30 minutes. The phosphorylation sites Thr58 and Ser62 exert opposite effects on the control of c-Myc degradation through the ubiquitin-proteasome pathway (Flinn et al, 1998; Sears et al, 2000; Amati, 2004; Herbst et al, 2004; Welcker et al, 2004; Yeh et al, 2004). Recent data indicate that the stability of c-Myc is regulated by different sequence elements, i.e. the N-terminal "degron" that signals Myc ubiquitination and degradation, and the C-terminal "stabilon" that promotes its sequestration and stabilization into a subnuclear compartment (Herbst et al, 2004).

The N-terminal domain of c-Myc, which is essential for transcriptional and transforming activity, binds to a -tubulin (Alexandrova et al, 1995) and is released from it during mitosis to facilitate microtubule disassembly. The release of c-Myc from a -tubulin is regulated by c-Myc phosphorylation state (Noguchi et al, 1999; Gregory and Hann, 2000; Niklinski et al, 2000). c-Myc protein shows a predominant localization in the cytoplasm of interphase cells, while in proliferating cells its nuclear distribution is similar to that of some ribonucleoprotein (RNP)-containing structures (Spector et al, 1987), or is confined to large amorphous nuclear globules (Henriksson et al, 1988; Koskinen et al, 1991). The existence of a dynamic modification of c-Myc is suggested by the competition of phosphorylation and glycosylation for the same site, i.e. Thr58 (Kamemura et al, 2002).

The search for the precise intracellular localization of c-Myc in tumor cells, where its degradation is deregulated with a resulting abnormal stability of the protein in the nucleus (Flinn et al, 1998; Salghetti et al, 1999; Gregory and Hann, 2000; Niklinski et al, 2000; Herbst et al, 2004), revealed that phosphorylated c-Myc accumulates in the nucleus of tumor cells. Phosphorylated c-Myc is distributed in the form of spots of different sizes throughout the nucleus and in the nucleolus (Soldani et al, 2002), where c-myc transcripts were described (Bond and Wold, 1993). As clearly demonstrated in HeLa cells (Soldani et al, 2002), phosphorylated c-Myc does accumulate in large amorphous globules (Henriksson et al, 1998) and its distribution pattern is not reminiscent of the distribution of non-nucleolar RNP-containing structures, as reported by Spector et al (1987). Remarkably, in tumor cells treated with the antimitotic drug paclitaxel, the immunolabeling for phosphorylated c-Myc changed, and became more diffused throughout the nucleoplasm (Bottone et al, 2003; Supino et al, unpublished observations). A typical example of the nuclear distribution of phosphorylated c-Myc in tumor cells is shown in Figure 3.

The most common alteration affecting c-myc in human tumors is gene amplification (Nesbit et al, 1999), which can range from a single gene duplication to hundreds of copies. Many experiments based on the enforced expression of an exogenously introduced c-myc gene provided the evidence that c-myc amplification could sensitize tumor cells to apoptosis. The pro-apoptotic role for c-myc has been first shown in serum-starved primary or immortalized fibroblasts (Evan et al, 1992; Fanidi et al, 1992) and in IL-3-dependent myeloid cells upon withdrawal of the cytokine (Askew et al, 1991) and this role was further confirmed (Alarcon et al, 1996; Dong et al, 1977; Rupnow et al, 1998). Promising results have been obtained by Peltenburg et al (2004), who demonstrated that the stable transfection of IGR39D melanoma cells with c-myc causes a sensitization of tumor cells toward apoptosis.

Although it is well established that apoptosis can be induced by the enforced expression of exogenously introduced c-myc genes in several experimental systems, it is interesting to investigate whether constitutive overexpression of the resident c-myc gene in tumor cells is sufficient to induce apoptosis. A positive correlation between endogenous high level of c-myc and apoptosis propensity was found in lymphoblastic leukemic CEM cells, which harbor constitutive activation of c-myc and undergo serum starvation-induced apoptosis (Tiberio et al, 2001).

We addressed this question by examining the effect of different apoptogenic stimuli on tumorigenic and non-tumorigenic clones isolated from the SW613-S human

Figure 3. Nuclear localization of phosphorylated c-Myc in HeLa cells. Immunofluorescence experiments were carried out according to Bottone et al (2003). Red fluorescence: a -tubulin; green fluorescence: phosphorylated c-Myc.

colon carcinoma cell line. 12A1 cells (tumorigenic clone) harbor an endogenous high level of amplification of the c-myc gene, whereas B3 cells (non-tumorigenic clone) have a small number of copies of this gene (Lavialle et al, 1988). We found that only cells with endogenous c-myc overexpression activate the apoptotic machinery in response to serum deprivation (Donzelli et al, 1999) and after the treatment with etoposide, doxorubicin and vitamin D₃, which induce Fas-mediated apoptosis (Gorrini et al, 2003). The low levels of c-myc expression present in SW613-B3 cells were unable to activate Fas-mediated apoptosis, thus suggesting that only a high c-myc expression can bypass the lack of Fas receptor. Apoptosis driven by DNA damage and long term-culture was independent of c-myc expression (Gorrini et al, 2003). The same experimental system was used to define the effect of c-myc amplification on the response to the antimitotic drug paclitaxel. A high c-myc amplification level potentiates paclitaxel cytotoxicity, confers a multinucleated phenotype and promotes apoptosis to a high extent, thus suggesting that c-myc expression level is relevant in modulating the cellular responses to paclitaxel (Bottone et al, 2003).

In conclusion, the overexpression of c-myc could be a strategy for therapeutic applications, possibly by modulating myc levels, thus sensitising tumor cells to therapy. As an example of the clinical potential of the analysis of the c-myc expression level in tumors, recent data obtained on patients with ovarian cancer suggest that a high c-myc expression level could improve the chemotherapy response (Iba et al, 2004).

The identification of genes that are important for the development and maintenance of malignant phenotype opened new perspectives for eventually inducing a reversion to normal phenotype. In this view, disease-associated proteins can be targets of a selective therapy that would lead to less toxic side effects than the conventional, often cytotoxic, therapeutic treatment. In fact, the main limitation to conventional cancer chemotherapy derives from the lack of specificity of the drugs, and from pharmacokinetic and manufacturing problems, which can lead to systemic, and organ toxicity. This impairs the use of high-dose intensity therapy, giving rise to a high rate of tumor relapse. The identification of fundamental genetic differences between malignant and normal cells resulting, for example, from activated oncogenes and inactivated tumor suppressor genes, has made it possible to consider such genes as specific targets for antitumor therapy. In this respect, many genes have been selected for antisense therapy, including HER-2/neu, PKA, TGF-a , EGFR, TGF-ß, IGFIR, P12, MDM2, BRCA, Bcl-2, ER, VEGF, MDR, ferritin, transferrin receptor, IRE, c-fos, HSP27, c-myc, c-raf and metallothioneins. Similar effects can be obtained with triple helix-forming oligonucleotides (TFOs) that are synthesized as to bind with a high affinity and specificity to double stranded DNA.

Additional approaches to modify c-myc expression consist of peptides, PNA (peptide nucleic acids) and siRNA (Cutrona et al, 2000; Hosono et al, 2004). Remarkably, it has been shown that c-Myc expression can be lowered by affecting the stabilization of a G-quadruplex structure present in the c-myc promoter (Grand et al, 2004).

AS-ODN are able to inhibit specifically the synthesis of a particular protein by binding to protein-encoding RNA, thereby preventing RNA function and thus inhibiting the action of the gene. Antisense therapy should correct the mutations and abnormal expression of genes of tumor cells by decreasing their expression, inducing RNA degradation, and causing a premature termination of RNA transcription (Head et al, 2002). Oligonucleotides (ODNs) are short pieces of DNA; their size ranges generally from 18 to 21 nucleotides. They hybridize to a specific target mRNA and their action can be mediated by the cleavage of the target DNA or by blocking the translation of RNA. In the first case, once the AS-ODN is bound to the specific RNA target, cellular RNase H cleaves the RNA/ODN complex, cleaving the RNA strand and releasing the ODN which can bind another specific RNA strand. Alternatively, ODNs ribozymes can be designed to hybridize and cleave the target RNA, thus to sterically bind RNA, with a resulting arrest of translation process.

TFOs are synthesized as to bind with a high affinity and specificity to the purine strand in the major groove of homopurine-homopyrimidine sequences in double stranded DNA. They can bind to DNA by parallel or anti-parallel orientation. TFOs directed against the purine-rich tracts of gene promoter regions are able to selectively reduce the transcription and the expression of target genes, by blocking binding of transcriptional activators and/or formation of initiation complexes. TFOs can be used to mediate site-specific genome modification. Indeed, TFOs are effective by binding as third strands with sequence specificity and the resulting triple helices, or TFO-mutagen complexes, are able to provoke repair and recombination (Faruqi et al, 2000), leading to directed mutagenesis, recombination, and, potentially, gene correction. TFO against p53, c-myc, bcl-2, HER/neu EGFR, etc have been successfully synthesized (Thomas et al, 1995; Basye et al, 2001; Shen et al, 2003; Re et al, 2004).

TFOs directed to regulatory sequences in the c-myc gene have been shown to inhibit transcription factor binding and transcription in vitro as well as promoter activity and gene expression in HeLa and MCF-7 cells (Postel et al, 1991; Thomas et al, 1995; Kim et al, 1998). Moreover, GT-rich TFOs directed to a sequence near the P2 promoter were particularly effective in inhibiting c-myc expression in leukemic and cancer cells (Catapano et al, 2000; McGuffie et al, 2000); daunomycin-conjugated GT-TFOs showed an increased stability of triple-helix and thus a higher activity of the TFO in human prostate (DU145) and breast cancer (MCF-7 and MDA-MB-231) cells (Carbone et al, 2004).

Although ODNs are under clinical investigation in different diseases, the majority of them are exploited against cancer for which this form of molecular therapeutics seems particularly suitable (Biroccio et al, 2003). ODNs are systemically administered and their toxicities, similar for all compounds, include thrombocytopenia, hypotension, fever and fatigue. AS-ODNs against c-myc are currently in phase I study in humans. The lack of toxicity together with the results obtained in a large amount of preclinical results (Iversen et al, 2003; Bayes et al, 2004) support their temptative therapeutic use.

It should be remembered that many other antisense approaches, including for example antisenses against BCL2, XIAP, PKA type I, EGFR, COX-2 inhibitors, gave, alone or in combination with antitumor agents, preclinical encouraging results in patients with advanced solid malignancies (Mani et al, 2003). Indeed this treatment is well tolerated and it is now in Phase III trials on chronic lymphocytic leukaemia, non-small-cell lung cancer, advanced malignant melanoma, multiple myeloma and prostate carcinoma (Hu and Kavanagh, 2003; Kim et al, 2004). Moreover, the effectiveness also of the oral administration of this kind of treatment makes this strategy very promising in cancer therapy (Tortora and Ciardiello, 2003).

Figure 4. Factors that can limit the use of antisense oligonucleotides (AS-ODN) or triple helix forming oligonucleotides (TFO). Sites 1-3 define where the factors reported in the respective boxes can interfere.

residence time of the oligonucleotides on the target and to increase their stability was to modify ODNs and TFOs as phosphorothioate oligonucleotides, which show a binding affinity similar to that of the phosphodiester oligonucleotide. A marked inhibition of c-myc transcription in HeLa cells has been demonstrated (Kim et al, 1998). Advantages in the affinity and the half-life of the binding of TFO to DNA were taken by the daunomycin-conjugated TFO; with this approach c-myc-targeted TFO showed a high stability and biological activity in mammary and prostate carcinoma cells (Carbone et al, 2004).

The oncogene c-myc plays essential roles in controlling cell cycle and proliferation, differentiation, tumorigenesis and apoptosis. For its crucial involvement in the development of cancer as well as in driving tumor cells to apoptosis, c-myc is a good candidate for the development of strategies aimed at modulating its activity in tumor cells.

In this respect, it is generally assumed that an increased level of c-myc could confer a propensity to apoptosis to a tumor cell, which is effective in potentiating the effects of clinical treatments. Even if this pro-apoptotic effect could be cell- and drug-dependent, promising results have been obtained in c-myc-overexpressing tumor cells derived from therapy-resistant tumors, such as melanomas and colon carcinomas.

An opposite strategy to face tumor development is the inhibition of the activity of factors that control cell proliferation and transformation, including c-myc. This goal is mainly achievable by the use of AS-ODN or TFO. The increasing amount of preclinical data on the effect of AS-ODN to c-myc encourages their temptative therapeutic use. However, potential limitation to gene-targeted therapies may exist, e.g. the development of resistant tumor cell populations that lose their sensitivity toward c-myc inhibition over time. In addition, since c-myc is a factor involved in determining the fate of normal cells and tissues, the side effects of its inactivation have to be considered.

In parallel with the antisense approach, the use of PNA and siRNA could provide an alternative way of down-regulating c-myc. The modulation of the functional interaction of c-Myc with its partners as well as the development of molecular tools to block the c-myc promoter could contribute to improve the anticancer therapy. Further in vitro experiments on different cancer cell lines will help in developing clinical trials aimed at obtaining a beneficial up- and down-regulation of c-myc in human tumors.

The research at the laboratory of RS and AIS is supported respectively by AIRC (Associazione Italiana Ricerca sul Cancro) and MIUR (FIRB Project RBNE0132MY).