The tumor suppressor functions attributed to Fhit protein (cell cycle and apoptosis regulator) are often observed when its expression is restored in cells that have lost sensitivity to damage signals. In this respect, several lines of evidence point to a role for Fhit as a sensor that can restore the sensitivity of cancer cells to external and/or internal stimuli. It has been hypothesized that modulation of Fhit expression in normal cells does not produce gross phenotypic changes, but rather results in a different behavior under particular conditions (e.g., UV or mitomycin C treatment). Consistent with this hypothesis is the phenotype of Fhit −/− mice, which are viable and grow normally, but have a higher incidence of spontaneous and especially carcinogen-induced tumors compared to the normal littermates.
Insights from mouse models
Several studies have been conducted in mouse models, since the FHIT murine locus is highly similar to the human locus (Glover et al., 1998), encompasses the chromosomal fragile site, and is susceptible to DNA breaks after exposure to carcinogens that inactivate the FHIT gene in preneoplastic and neoplastic lesions.
In vivo models using mouse strains with one or both inactivated alleles have been established to investigate the events required for cancer development (Table 1). Fong et al. (2000) compared the susceptibility of Fhit +/+ and Fhit +/− mice to tumor formation induced by NMBA, a carcinogen that produces esophageal and forestomach tumors. The study showed that after eight doses of carcinogen, 100% of Fhit hemizygous mice developed tumors of the gastrointestinal tract as compared to only 25% of Fhit +/+ mice. In addition, 60% of Fhit +/− mice also developed sebaceous tumors with a phenotype similar to that of tumors in patients with Muir–Torre syndrome, suggesting that Fhit may be a target of damage in a fraction of mismatch repair-deficient cancers. Of interest, the development of NMBA-induced tumors in these mice could be prevented by administration of Fhit-expressing viral vectors (Dumon et al., 2001a).
Table 1. Tumors incidence in Fhit-deficient mice
|Genotype||Tumors incidence (%)|
|Spontaneous||Forestomach (NMBA eight doses)||Forestomach (NMBA single dose)||Bladder (BBN)||Lung (cross with Vhl +/−)|
|Fujishita et al., 2004||Zanesi et al., 2001||Fong et al., 2000||Zanesi et al., 2001||Vecchione et al., 2004||Zanesi et al., 2005|
|FHIT +/−||60%||52.9%||100%* (60% MTS-like)||78.3%||46%||ND|
A subsequent study in which mice with one, both, or neither intact FHIT allele were treated with a single dose of NMBA (Zanesi et al., 2001) found that more than 75% of mice Fhit +/− and Fhit −/− developed tumors compared to 8% of wild-type mice. Additionally, the authors found that more than 50% of heterozygous and nullizygous mice developed spontaneous tumors compared to only 8% of Fhit +/+ mice. Other authors have also demonstrated that untreated Fhit +/− and Fhit −/− knockout mice have a higher incidence of spontaneous tumors than do wild-type mice (Fujishita et al., 2004). These studies provided evidence that the loss of one Fhit allele had the same effect on tumor development as the loss of both alleles, suggesting that Fhit might be haplo-insufficient for tumor suppression.
The situation seems to be more complicated when other organs are considered. Recently, Vecchione et al. (2004)) examined the role of Fhit in the development of bladder cancer using FHIT knockout mice treated with BBN, a potent carcinogen that induces bladder tumors. The authors observed that although 76% of Fhit +/+ mice developed hyperplasia and mild dysplasia, only 8% of those mice showed the invasive carcinoma present in 46% of Fhit +/− and 28% of Fhit −/− mice. However, Fhit +/− mice appeared to be more susceptible to carcinogens compared to Fhit−/− mice, and the authors raise the possibility that Fhit plays a role in modulating unknown partners in bladder carcinogenesis.
Several studies have reported the loss of Fhit expression in lung preneoplastic and neoplastic lesions (Sozzi et al., 1998; Tseng et al., 1999), which frequently exhibit alterations on chromosome 3p, the site of FHIT and other tumor suppressor genes. To analyze the potential cooperation in tumor suppression by different genes on 3p, the incidence of spontaneous and induced lung tumors was evaluated in a cross between mice deficient for FHIT and Vhl, a 3p26-p25 gene frequently altered in lung cancer. The authors observed that 44% of Fhit −/− Vhl +/− mice developed spontaneous lung tumors by 2 years of age compared to none of the single Vhl +/− or Fhit −/− mice. In addition, double-deficient mice had a tumor incidence of 100% after carcinogen treatment. This incidence was higher than that observed for Fhit −/− mice (40%), suggesting that FHIT-deficient mice required further alteration for lung cancer progression. However, the number of mice analyzed in that study was relatively small, and the relationship between FHIT and Vhl requires further investigation (Zanesi et al., 2005).
Cell lines derived from the FHIT mouse model are also proving to be valuable tools in dissecting the molecular pathways involving Fhit. Ottey et al. (2004) observed reduced levels of apoptosis and enhanced clonogenic survival in Fhit −/− cells compared to Fhit +/+ cells after exposure to mitomycin C and UV-C treatment. In addition, Fhit −/− cells had a strong DNA checkpoint, regulated by an over-activation of the ATR/CHK1 pathway, which contributed to the radio-resistance and thus cell survival (Hu et al., 2005a). In particular, Fhit and Chk1 appear to have opposing roles in homologous recombination repair (Hu et al., 2005b), with loss of Fhit expression leading to enhanced repair activity and possibly to increased survival of cells with increased mutation burdens (Fig. 2).
Figure 2. FHIT and damage activated pathways. Schematic representation of how loss of Fhit expression could contribute to carcinogenesis by providing survival advantage of damaged cells.
Download figure to PowerPoint
Both in vitro and animal model studies continue to provide valuable pieces of the puzzle of Fhit's role in tumorigenesis and are instrumental in developing novel therapeutic tools in cancer prevention and treatment (Ishii et al., 2004).
Fhit regulation of the cell cycle and apoptosis
Evidence obtained mainly through analyses of cells isolated from the Fhit knockout mouse and through transduction experiments in cancer cell lines supports the involvement of the Fhit protein in cell cycle and apoptosis regulation. Although Fhit −/− and Fhit +/+ cells isolated from mouse kidneys do not differ in their proliferation properties, the Fhit-negative cells reportedly show an increase in the S-phase of the cell cycle (+ 10% of cells) and a decrease in the G1-phase (− 10% of cells) compared to the normal counterpart, suggesting that the S and G2 checkpoints might be over-activated (Ottey et al., 2004). Consistent with this notion is the apparent contribution of a strong S-phase checkpoint in Fhit-negative cells to the UV resistance of these cells, slowing damaged cells before an apoptotic response is triggered. It seems, therefore, that in normal cells Fhit could participate in the regulation of the S/G2/M transition, possibly by facilitating the entrance into apoptosis of cells with genomic alterations.
One of the first effects observed in transduction experiments using Fhit-negative lung cancer cells was a transient accumulation of cells in G2 phase, before the onset of an apoptotic effect resulting in the appearance of the sub-G0 peak and accumulation of the remaining cells in G1 and S phases (Roz et al., 2002).
In general, restoration of Fhit expression in cancer cells reduces the proliferation rate (Ishii et al., 2001a), although not all lines tested respond to Fhit restoration treatment (Werner et al., 2000). This non-responsiveness might rest in tissue-specific differences, in additional underlying genetic alterations, or in factors related to expression levels of either the transduced or the residual endogenous protein. In fact, an accurate analysis performed using an inducible system showed that the anti-proliferative effect of Fhit is exquisitely dose-dependent (Cavazzoni et al., 2004).
At the molecular level, the growth suppressive properties of Fhit have been linked to upmodulation of the cell cycle regulator p21waf1, a potent and tightly binding inhibitor of cyclin-dependent kinases. This link has been demonstrated in different experimental models at both the RNA and protein levels (Sard et al., 1999), and has been shown to be p53-independent (Cavazzoni et al., 2004). The relationship between Fhit and p53 pathways has also been investigated extensively in light of the high frequency of alterations of both genes in human cancers, and has revealed a synergistic oncosuppressor activity (Nishizaki et al., 2004). Although some reports suggest that Fhit does not require a functional p53 for its activity, since even p53-negative cells showed reduced growth and evidence of apoptosis after Fhit transduction (Ji et al., 1999; Roz et al., 2002), the coordinated expression of the two proteins has a very potent anti-tumor effect. While this synergistic effect might be due to stabilization of p53 related to Fhit-mediated downregulation of MDM2 (Nishizaki et al., 2004), it more likely rests in the convergence of the two pathways on a common mediator such as p21waf1 (Cavazzoni et al., unpublished data).
Other clues to the effects of Fhit on the cell cycle come from microarray experiments where the majority of regulated transcripts after FHIT restoration were represented by genes involved in mitotic control; coordinated downregulation of kinesin family members (KNSL1, KNSL6), DNA replication factors (MCM2, MCM5), and proteins involved in spindle assembly checkpoints (BUB1, KNTC2) has been observed (Roz et al., 2003).
The onset of apoptosis after Fhit expression restoration has been consistently reported in many studies (Ishii et al., 2001a). Three main characteristics appear to typify Fhit-induced apoptosis: (i) slow onset, requiring long-lasting and sustained Fhit expression; (ii) significant enhancement by other apoptotic stimuli; and (iii) general absence in cells with normal Fhit expression. Thus, a likely scenario is that Fhit acts to lower the apoptotic threshold of damaged (cancer) cells. A putative sensor role of Fhit would explain the time- and dose-dependence of the response, the synergy with other stimuli, and the absence of any perturbation in normal cells by overexpression of this protein.
Several mediators have been implicated in Fhit-mediated apoptosis, including activation of different caspases (-8, -9, -3, and -2), cleavage of multiple substrates (PARP, β-catenin, Bid), and loss of mitochondrial potential (Dumon et al., 2001b; Ishii et al., 2001b; Roz et al., 2002). Consistent with these data, caspase inhibitors have been shown to prevent apoptosis after Fhit restoration (Ishii et al., 2001b).
Two main apoptotic pathways have been described, the “extrinsic” (cytoplasmic) pathway, which is strongly linked to signals from membrane receptors and headed by caspase-8, and the “intrinsic” (mitochondrial) pathway, regulated mainly by Bcl-2 family members at the mitochondrial level and reliant on caspase-9 activation. To understand how Fhit relates to apoptosis, the connection of the two pathways at the cellular level by several mediators must be considered. Caspase-8-mediated activation of the mitochondrial pro-apoptotic factor Bid represents one of the best-known links. In cells where the two pathways are strongly linked, observable involvement of mediators of both pathways in the late stages of the apoptotic response is expected. However, lung cancer cells with a blocked mitochondrial apoptotic program due to Bcl-2 or Bcl-X(L) overexpression still exhibit apoptosis, despite the absence of cytochrome-c release from the mitochondria (Roz et al., 2004), indicating that this amplification step is not necessary for Fhit activity. This is consistent with the reported activation of caspase-8 in the early phases of Fhit-induced apoptosis (Roz et al., 2002) and suggests that Fhit exerts its apoptosis-facilitating function mainly at the cytoplasmic level.
These observations have potential clinical and therapeutic implications since knowledge of the apoptotic pathways triggered by different agents can provide valuable information about chemo-resistance mechanisms and the possible synergy or antagonism among different treatments. In particular, not only are Fhit −/− cells reportedly mitomycin C- and UV-C-resistant (Ottey et al., 2004), but also Fhit was recently shown to modulate sensitivity to cisplatin in lung cancer cells (Andriani et al., 2005).