This Article  • Abstract  • PDF  • Send to a friend  • Save in My Folder  • Save to citation manager  • Permissions  Citing Articles  • Citation map  • Citing articles on HighWire  • Citing articles on ISI (42)  • Contact me when this article is cited  Related Content  • Related article  • Similar articles in JAMA  Topic Collections  • Review  • Genetic Counseling/ Testing/ Therapy  • Alert me on articles by topic

Review of the Preclinical Studies

Hideshi Ishii, MD,PhD; Kristoffel R. Dumon, MD; Andrea Vecchione, MD; Louise Y. Y. Fong, PhD; Raffaele Baffa, MD; Kay Huebner, PhD; Carlo M. Croce, MD

JAMA. 2001;286:2441-2449.

ABSTRACT
Context  The fragile histidine triad gene (FHIT) encompasses a human common fragile site, FRA3B, that is susceptible to environmental carcinogens. Deletion and inactivation of FHIT have been seen in a number of human premalignant and malignant lesions.

Objective  To review and evaluate preclinical studies of cancer therapy using the FHIT tumor suppressor gene and related studies involving Fhit protein expression.

Data Sources  A MEDLINE search of articles published from 1996 to June 2001 was performed; article reference lists were used to retrieve additional relevant articles.

Study Selection  Immunohistochemical studies of primary tumors or relevant lesions were selected to evaluate Fhit expression in premalignant or malignant stages. Preclinical studies on antitumorigenic or therapeutic introduction of FHIT were reviewed for the effects of exogenous Fhit expression. For the immunohistochemical analyses, 26 studies were included that analyzed at least 15 cases of a single type of tumor. For precancerous lesions, 9 studies were included that analyzed at least 4 cases. For studies of FHIT introduction, 9 published studies were included.

Data Extraction  Using primary data from each of the studies, we assessed the rationale and potential contribution of FHIT cancer therapy. Data was independently abstracted by 2 authors and study quality was assessed by 2 other authors.

Data Synthesis  Overall, 60% (1162/1948 cases) of primary tumors showed absent or markedly reduced Fhit protein expression in cancer cells. Studies of preneoplastic lesions or early-stage cancer showed absence or marked reduction of Fhit protein expression in 0% to 93% of samples (overall, 31% [127/408 cases]). Preclinical studies using 26 cancer-derived cell lines from human lung, head and neck, esophageal, gastric, cervical, pancreatic, and kidney cancers, showed that reintroduction of FHIT resulted in inhibition of in vitro tumor cell growth or of in vivo tumorigenicity in 17 (57%) of 30 cell line experiments. Model systems for human preventive cancer therapy suggested that oral introduction of viral vector-mediated FHIT into Fhit-deficient mice may prevent carcinogen-induced tumor development in some cases.

Conclusion  These findings show that FHIT gene therapy may potentially be clinically useful for treatment of cancer and also prevention of carcinogen-induced tumor development, suggesting a rationale for further research involving FHIT introduction.

INTRODUCTION

Jump to Section  • Top  • Introduction  • Methods  • Results  • Comment  • Author information  • References

Considerable evidence has accumulated suggesting that cancer has a genetic origin, based on the development of genomic alterations, such as deletions, translocations, mutations, and/or methylation in critical genes for homeostasis of cellular functions, including cell cycle control, cell survival, and DNA replication.^1-3 Genomic alterations that inactivate suppressor genes or activate oncogenes could relieve cells from growth constraints, allowing tumor growth.^1-3 Current studies of the molecular basis of cancer provide a strong rationale to consider gene therapy approaches to cancer as a complement to other cancer therapies.^4-5 Restoration of the function of a single or several pivotal suppressor gene products appears sufficient to mediate antitumor effects that could be clinically significant.^4-5 Tumor suppressor genes, such as TP53 or p16(MTS-1/CDKN2/INK4A), which encode proteins regarded as key effectors of cellular senescence and control exit from the G1 phase of cell cycle, are inactivated in human cancers.^6-8 In vitro and in vivo studies have shown that restoring tumor suppressor function in cancer cells leads to suppression of cancer cell growth through cell cycle arrest or apoptosis.^6-8

Human chromosomal fragile sites are specific chromosomal regions that, under appropriate culture conditions, including exposure to aphidicolin, an inhibitor of DNA replication, exhibit the tendency to form chromosome or chromatid gaps.⁹ These gaps are visible in metaphase chromosomes and represent decondensation of the chromatin structure, some of which may represent breaks in the DNA. A number of aphidicolin-induced specific breaks and viral DNA integration sites have been identified at the carcinogen-sensitive common fragile site, the FRA3B locus.^10-12 The fragile histidine triad (FHIT) gene, isolated through positional cloning, encompasses the FRA3B locus at chromosome 3p14.2, a region involved in hemizygous or homozygous deletions in many types of human tumors.¹³ The FHIT gene spans approximately 1.8 megabase (Mb) (T. Shiraishi, MD, PhD, and C.M.C., unpublished data, April 2001) and is the second largest human gene after the 2.4-Mb dystrophin gene.¹⁴ The size of the FHIT messenger RNA (mRNA), however, is 1.1 kilobase (kb).¹³ The aphidicolin-induced fragile region extends at least from exon 2 to exon 10 of the FHIT gene, but the majority of DNA breaks occur in FHIT introns 4 and 5, flanking the first FHIT coding exon (exon 5); not surprisingly, exon 5 is frequently the location of biallelic deletion in human cancer and cancer-derived cell lines.^{13, 15-16}

Numerous studies have shown genomic alterations at the FHIT locus such as deletions and translocations and loss of Fhit protein expression in a number of premalignant and malignant lesions of lung, cervix, esophagus, breast, and other organs (Figure 1),^17-21 as reviewed in this article, representing progress toward a better understanding of the mechanism of gene inactivation. While point mutations within the FHIT gene are rare,^22-26 deletions are common,^15-16 and to a lesser extent, silencing by methylation is involved.²⁷ A germline FHIT allele was found to be interrupted by a t(3;8) chromosome translocation in familial clear cell renal carcinoma.^{13, 15} A patched-related gene, TRC8, at chromosome 8q24, was found to fuse to the 5' noncoding exon of FHIT in the t(3;8) translocation, resulting in disruption of TRC8 within the sterol-sensing domain.²⁸ The study showed a tumor-specific alteration of TRC8 in 1 of 32 sporadic renal carcinomas examined (3%). Another study showed that the FHIT gene had fused to the high mobility group (HMGIC) gene at chromosome 12q15 in a primary pleomorphic adenoma.²⁹ In the translocation, the open reading frame of each FHIT and HMGIC gene was disrupted, resulting in abnormal fusion products, Fhit-Hmbci and Hmbci-Fhit, in the benign tumor, suggesting that 2 tumor-associated genes are concurrently important in development of adenoma of the salivary glands.²⁹

View larger version (56K): [in this window] [in a new window] Figure. Immunohistochemical Analyses of Fhit Protein Human diffuse-type gastric cancer (A) and human transitional cell bladder carcinoma (B) are depicted. Black arrowheads indicate tumor with negative Fhit staining, while white arrowheads indicate Fhit-positive normal epithelium. The sections were immunostained with anti-Fhit antibody.

These findings suggest that alterations and inactivation of the FHIT gene contribute to tumor development. Nucleotide sequence analysis of the FHIT locus in cancer cell lines has indicated that long interspersed nuclear elements and long terminal repeat sequences are often involved in homologous recombination at the deletion end points, resulting in loss of portions of the FHIT gene.^30-31 Presumably because FHIT encompasses the carcinogen-sensitive fragile region, it is especially susceptible to damage from environmental carcinogens, such as in cigarette smoke,^16-17 leading to clonal expansion of Fhit-negative cells.

Stable FHIT reintroduction and expression in Fhit-negative lung, gastric, and renal cancer cells have resulted in the suppression of tumor cell growth,^32-34 at least in part due to induction of apoptosis,³³ showing the FHIT tumor suppressor function. It has also been shown that reintroduction of FHIT and expression of Fhit protein by viral vector-mediated transduction in lung, head and neck, esophageal, and pancreatic cancer cell lines can cause apoptosis and inhibition of tumorigenicity.^35-37 Recently, adenoviral or adenoassociated viral-mediated FHIT expression was shown to inhibit tumor development after oral administration to cancer-prone Fhit knockout mice.³⁸ The adenoviral vector is derived from an adenovirus, whereas the adenoassociated virus can propagate in the presence of adenovirus as a helper virus. Adenoviral DNA is replicated as an episomal but the adenoassociated virus can be integrated.³⁹ These data suggest a rationale for further investigation of FHIT gene therapy as part of clinical trials designed to test new approaches to cancer therapy or prevention.

METHODS

Jump to Section  • Top  • Introduction  • Methods  • Results  • Comment  • Author information  • References

A MEDLINE search was performed with use of the following keywords: FHIT, fragile histidine triad gene, neoplasm, lung cancer, head and neck cancer, esophageal cancer, gastric cancer, colorectal cancer, pancreatic cancer, breast cancer, cervical cancer, endometrial carcinoma, kidney cancer, prostate cancer, bladder cancer, transfection, adenoviral vector, adenoassociated viral vector, protein expression, immunohistochemistry, carcinogen, and/or N-nitrosomethylbenzylamine (NMBA). Articles published from 1996 to June 2001 were accessed and reviewed, and using references contained therein, additional relevant publications were retrieved. In the immunohistochemistry studies of cancer, 26 studies were included that analyzed at least 15 cases of a single type of tumor. Nine studies were included that analyzed at least 4 cases for precancerous lesions. For studies of FHIT introduction, 9 published studies were included. Data were independently abstracted by 2 authors blinded to the journal, authors, and funding sources. Study quality was assessed by 2 other authors, blinded to the journal, title, authors, introduction, and discussion. Criteria for assessment of quality included the number of cases and the quality of the representation of data.

RESULTS

Jump to Section  • Top  • Introduction  • Methods  • Results  • Comment  • Author information  • References

Fhit Protein Expression in Primary Tumors and Premalignant Lesions

Immunohistochemical studies of primary tumors indicated that Fhit protein expression in cancer cells was absent or markedly reduced in many different cancers compared with noncancerous tissue (Table 1). A relatively large-scale study showed that Fhit protein was lost or reduced in 73% (345/474 cases) of non–small cell lung cancer, including 87% of cases of squamous cell carcinoma of lung (202/233 cases) with marked reduction in immunohistochemical staining (0%-10% staining).¹⁷ Other studies, relatively small in scale, showed that Fhit protein was lost or reduced in 70% to 79% of esophageal squamous cell carcinoma,^{19, 40} 23% to 68% of head and neck squamous cell carcinoma,^41-43 62% to 67% of gastric cancer,^44-45 27% to 72% of breast cancer,^{20, 46} 40% to 71% of cervical cancer,^{18, 26, 47-48} 37% to 48% of endometrial cancer,^49-50 or 51% to 78% of clear cell renal carcinoma.^51-52 Other relatively small-scale studies showed loss or reduction of Fhit protein in 44% to 81% of colorectal, pancreatic, prostate, or bladder cancer.^53-56 Overall, 60% (1162/1948 cases) of different primary tumors showed loss or reduction of Fhit protein expression in cancer cells. Criteria for determining level of protein expression vary among studies. For example, experimental conditions for the immunohistochemistry studies, involving expression of antigen, dilution factor for antibodies, and time and temperature of incubation, may vary. However, each report can be evaluated in terms of comparing degree of staining of cancer cells with that of normal tissue as an internal control, regarding the protein expression of interest.

View this table: [in this window] [in a new window] Table 1. Loss or Reduction of Fhit Protein in Human Cancer*

Studies of preneoplastic lesions or early-stage cancer showed absence or marked reduction of Fhit protein expression in 0% to 93% of precancerous lesions of various tissues (Table 2). Fhit expression was lost in a total of 60% (21/35 cases) or 93% (42/45 cases) at an early stage of esophageal or lung carcinogenesis, respectively.^{17, 19, 61} Overall, 31% (127/408 cases) showed loss or marked reduction of Fhit protein expression.

View this table: [in this window] [in a new window] Table 2. Fhit Protein Expression in Preneoplastic Lesions or Early Stage Cancer*

Previous studies have shown that loss of Fhit expression in lung cancer can be related to smoking.^{17, 57-58,63-64} Absence or marked reduction of Fhit protein is significantly higher in tumors of smokers (40%-75% cases) than in those of nonsmokers (16%-39% cases),^{17, 58} and is an independent event^{17, 59, 65-66} and occurs more frequently than p53 alterations.^{17, 66} These findings suggest use of Fhit loss as an early detection marker and FHIT alterations as targets for early gene therapy.¹⁷ An association between asbestos exposure and smoking duration and FHIT exon loss was detected through analysis of primary lung tumors, indicating that carcinogenic exposures may be a cause in the generation of FHIT alterations.⁶⁴ Tumors in nonsmoking subjects account for less than 10% of all lung cancers and show distinct histopathological features.⁶⁷

A significant correlation between heavy smoking and alcohol use and FHIT inactivation has been shown in esophageal cancer.¹⁹ It was shown that Fhit expression was lost in 32 of 46 esophageal carcinoma cases tested, in 8 of 12 with carcinoma in situ, and 9 of 21 with dysplasia.¹⁹ The relatively small study demonstrated that tumors from heavy users of both tobacco and alcohol show a significantly higher frequency of loss of Fhit expression than those of light users.¹⁹ Interestingly, the study showed that noncarcinomatous epithelium, corresponding to 46 esophageal carcinoma cases, displayed positive Fhit reactivity in all except 5 cases; these 5 patients had habits of heavy use of tobacco and alcohol.¹⁹ It has been shown that normal epithelium, including that of esophageal and stomach, expresses FHIT mRNA¹³ and Fhit protein.^{19, 44} The study suggested that loss of Fhit expression is an early event in the development of esophageal carcinoma and predisposing genetic changes have occurred even in normal-appearing epithelium in cases heavily exposed to environmental carcinogens. However, another study concluded that, although Fhit expression is associated with progression of esophageal carcinoma, it is not associated with smoking history⁴⁰; it also has been suggested that not only smoking but also dietary factors could be involved in esophageal carcinogenesis,⁶⁸ and that smoking may contribute to carcinogenesis in the digestive tract¹⁹ as well as respiratory organs.¹⁶ Studies of large panels of cancers will be needed to reconcile these varying observations.

Effect of FHIT Gene Introduction Into Cancer Cells

Experimental studies using 26 cancer-derived cell lines from human lung, head and neck, esophageal, gastric, cervical, pancreatic, and kidney cancers, showed that reintroduction of FHIT results in inhibition of tumor cell growth in vitro and/or tumorigenicity in vivo in 17 of 30 cell line experiments (57%) (Table 3). Endogenous (ie, derived from cell chromosomes) Fhit protein expression was assessed via immunoblot analysis and exogenous (ie, derived from FHIT introduction into cells from outside) Fhit protein was also determined via immunoblot analysis. It is difficult to determine the target population for clinical trials exactly and to estimate the potential effect in clinical trials, partially because of insufficient background information about populations of cell lines examined, and of the transduction and expression efficiency per cell. Among Fhit-negative cancer cell line experiments reported, tumor suppression was observed in 13 (62%) of 21 cases, whereas 4 (44%) of 9 Fhit-positive or weakly positive cancer cell line experiments reported showed substantial tumor suppression. Different pathways may be involved; in any case, preclinical studies of large panels of cancers will be necessary to completely define the spectrum of cancer susceptible to FHIT gene therapy.

View this table: [in this window] [in a new window] Table 3. Experimental Transduction of FHIT Gene Into Cancer Cells*

Studies of FHIT introduction into esophageal and renal cancer cells show that susceptibility to apoptosis or cell growth inhibition is not restricted to cancer cells with complete loss of endogenous Fhit expression and could be dependent on cell type.^{34, 36} It has been reported that cervical cancer cells stably expressing exogenous Fhit showed no significant alteration in cell growth.^69-70 When we expressed the Fhit protein in HeLa cells by adenoviral-FHIT³⁶ and adenoassociated viral FHIT infection (K.R.D. and C.M.C., unpublished data, April 2001), HeLa cells showed marked apoptosis in each experiment, suggesting that perhaps the threshold of Fhit expression necessary for biological effect was not reached in the earlier studies. The amounts of the Fhit protein expressed in cancer cells are approximately 0.5- to 5-fold in stable transfection and up to 50-fold in adenoviral transduction, relative to the endogenous Fhit protein level.³⁶ The adenoassociated viral transgene expression can be up to 10-fold the level of endogenous Fhit.³⁷ The expression level of the FHIT transgene and the route of gene delivery may not be crucial factors for induction of tumor suppression (Table 3).

After Siprashvili et al³² first showed tumor suppression with stable FHIT reintroduction and expression in Fhit-negative cancer cells, additional evidence of the tumor suppressor function of Fhit has been presented,^33-37 although not all studies have observed tumor suppression associated with exogenous Fhit expression.^{34, 69-70} Adenoviral or adenoassociated viral FHIT expression significantly inhibited cell growth of human lung cancer cells, head and neck carcinoma cells, esophageal cancer cells, and pancreatic cancer cells, but not of normal bronchial epithelial cells.^35-37 Although not all cell types have shown tumor suppression associated with exogenous Fhit expression, some overall observations are (1) that viral vector-mediated FHIT expression induces marked caspase-dependent apoptosis and alters cell cycle progression^35-37; (2) that tumorigenicity of cancer cells transduced by viral vector-mediated FHIT is suppressed in vivo^35-37; (3) that subcutaneous tumor growth of cancer cells in nude mice is significantly reduced after intratumoral injections of FHIT virus³⁵; (4) that Fhit expression enhances the susceptibility of pancreatic cancer cells to exogenous tumor necrosis factor³⁷; and (5) that FHIT gene transfer delays pancreatic tumor growth and prolongs survival in a murine in vivo model mimicking the human disease.³⁷ These observations indicate that exogenous Fhit expression can inhibit cancer cell growth in vitro and in vivo, at least in part through caspase-dependent apoptosis, suggesting that exogenous Fhit expression should be explored as a therapeutic strategy.

Several observations implicate Fhit in proapoptotic signaling. Our study with 7 esophageal cancer cell lines showed that proapoptotic molecules, such as caspase 9 and Bid, were cleaved in 2 cell lines, but were not or barely cleaved in 5 other cell lines after adenoviral-FHIT transduction, suggesting that both caspase 9 and Bid activation are required for onset of apoptosis.³⁶ Caspase 8 was cleaved in all 7 esophageal cancer cell lines, specifically after adenoviral-FHIT transduction, suggesting that caspase 8 is downstream of Fhit in a signaling pathway in all the esophageal cancer cells. When adenoviral-FHIT infected cells were cultured in medium with individual caspase inhibitors, protection from apoptosis was observed, suggesting that for full execution of Fhit-induced apoptosis, initiators including caspase 8 are required. These data prompt us to speculate that adenoviral-FHIT transduction results in activation, not only of the mitochondrial pathway, but also of the caspase 8 pathway, possibly amplified through Bid cleavage.⁷² This study suggests that FHIT apoptosis-resistant esophageal cancer cells could be refractory to the exogenous Fhit expression, at least in part because of nonresponsiveness of the mitochondrial or DNA cleavage/fragmentation pathway.

Although the human FHIT gene encodes a protein with 69% similarity to an Schizosaccharomyces pombe enzyme, diadenosine 5', 5'''-P¹, P⁴-tetraphosphate (Ap₄A) hydrolase, showing FHIT is a member of the histidine triad gene family,^{13, 73} the biological mechanism concerning tumor suppression is not fully understood. What is an actual trigger for Fhit-induced apoptosis in cancer cells? Fhit behaves in vitro as a typical dinucleotide 5',5'''P¹, P³-triphosphate (Ap₃A) hydrolase and mutation of a central histidine abolishes hydrolase activity,⁷⁴ but it has never been shown that this hydrolase activity has an in vivo biological function. Other approaches to understanding the biological function of Fhit showed association of the Fhit protein with microtubules⁷⁵ and with human ubiquitin-conjugating enzyme 9 (hUBC9), the yeast homologue of which is involved in the degradation of S- and M-phase cyclins.⁷⁶ The interaction between Fhit and microtubules or hUBC9 is independent of the Fhit enzymatic activity. Two studies showed that exogenous Fhit inhibits tumorigenicity in nude mice and that Ap₃A hydrolysis is not required for tumor suppression.^{32, 34} Wild-type Fhit and an Fhit protein resulting from a mutation involving a central histidine (having defective hydrolase activity) both suppressed tumorigenicity in nude mice, showing similar tumor suppressor activity.³² It has thus been hypothesized that the substrate-bound form of Fhit is the important biological signaling molecule.⁷³ We have explored gene expression in FHIT-induced esophageal cancer cells by differential display to assess gene expression profiles after adenoviral FHIT introduction (H.I. and C.M.C., unpublished data, April 2001). We analyzed 7 esophageal cancer cell lines, TE1, TE2, TE4, TE10, TE12, TE13, and TE14, before and after adenoviral FHIT introduction, and isolated several genes with increased expression in adenoviral FHIT-infected TE4 and TE14 cells, including CD9, profilin, cytochrome B, and mitochondrion genes. Genes with decreased expression included endothelial differentiation factor and mitochondrial creatine kinase genes. These genes may contribute to Fhit-related tumor suppression directly or indirectly, but molecular mechanisms for tumor suppression remain to be elucidated.

The Role of FHIT Gene Therapy in Preventing Tumor Development in Carcinogen-Treated Mice: A Preclinical Approach

The FHIT gene appears to be inactivated in an early stage of tumor development^17-20 and is deleted in a number of tumors.¹⁶ In addition to exploration of FHIT transduction for treatment of cancer, it is important to assess the contribution of FHIT allelic deletions in premalignant lesions and very early stages of cancer and to study the effect of reintroduction of FHIT in premalignant and early stage cancer.

Experiments in animal models show that Fhit knockout mice with 1 disrupted allele (Fhit + / - mice) are more susceptible to NMBA-induced upper gastrointestinal tract tumor development than wild-type mice (Fhit + / + ).⁷⁷ Of the heterozygous mice (Fhit + / - ), 100% had multiple tumors in esophagus, forestomach, and/or squamocolumnar junction (SCJ) after exposure to NMBA compared with 25% of mice with intact Fhit alleles (Fhit + / + ),⁷⁷ suggesting that hemizygous loss of Fhit results in greater susceptibility to carcinogen-induced tumor development. Similar to human distal esophagus, the mouse forestomach has a squamous epithelial lining. The mouse SCJ corresponds to the human esophago-gastric junction, commonly used as a model system for the human distal esophagus.³⁸ Fhit - / - mice are only slightly more sensitive to carcinogens than Fhit + / - mice, and both heterozygous and homozygous knockout mice exhibit increased frequencies of spontaneous tumors (N. Zanesi, PhD, and C.M.C., unpublished data, April 2001), suggesting that loss of 1 FHIT allele in an organ or tissue contributes to malignancy.

The most extensively used carcinogen for esophageal carcinoma induction in rodents is NMBA.³⁸ Epidemiological studies have shown that exposure to nitrosoamines such as NMBA is associated with the high incidence of esophageal cancer in northern China and in parts of Iran,^78-80 reinforcing the relevance of this animal model to human cancer. Benzaldehyde and an electrophilic agent methylating DNA are produced by NMBA when bioactivated, leading to the formation of a promutagenic adduct, O⁶-methylguanine.⁸¹

Recently we have been able to inhibit carcinogen-induced upper gastrointestinal tract tumor development by oral gene transfer of the FHIT gene, using adenoviral or adenoassociated viral vectors in heterozygous Fhit + / - knockout mice.³⁸ The control experiments showed that the adenoviral transgene expression was detected in the epithelial layer of the forestomach at the maximal level 3 days after vector administration, while adenoassociated transgene expression was undetectable at day 3, but became strongly positive in the epithelial cell layer and lamina propria of the forestomach and esophagus 14 days after vector administration. These findings are consistent with previous observations of the high efficiency of the adenoassociated viral vector in transducing endoluminal cells in the intestine.^82-83

In the experimental protocol, Fhit + / - mice (20 to 32 weeks old) received 6 doses of NMBA at 2 mg/kg of body weight during 3 weeks.³⁸ Ten days after the final NMBA administration, 1 group of 8 mice received 1 dose of adenoviral FHIT vector (10¹¹ plaque-forming units/mL, 100 µL); the second group (n = 8) received 1 dose of adenoassociated FHIT virus (10¹¹ viral particles/mL, 100 µL), and the third group (n = 6) received the same dose of adenoviral and adenoassociated FHIT virus in combination. Twelve mice did not receive any virus. All mice were euthanized for analysis 10 weeks after virus administration. In the forestomach, large visible tumors were seen in 11 (92%) of 12 mice given NMBA without FHIT virus; however, in FHIT-treated mice, tumors were seen in 4 (50%) of 8 adenoviral FHIT-treated mice (vs control, P = .11), in 3 (38%) of 8 adenoassociated viral FHIT-treated mice (P = .02), and 3 (50%) of 6 combined adenoviral and adenoassociated FHIT-treated mice (P = .08).³⁸ At the SCJ, tumors were seen in 8 (67%) of 12 control mice, in 0 (0%) of 8 adenoviral FHIT-treated mice (vs control, P = .005), in 0 (0%) of 8 adenoassociated viral FHIT-treated mice (P = .005), and in 1 (17%) of 6 combined adenoviral and adenoassociated FHIT-treated mice (P = .13).³⁸ One (8%) of 12 control mice developed tumors in the esophagus, whereas 22 (0%) FHIT-treated mice did not form esophageal tumors.³⁸ Pathological findings were observed in 100% of sections from control mice, including papilloma, focal hyperplasia, and invasive carcinoma, whereas the 3 groups of FHIT-treated mice showed reduction in pathological findings.³⁸ An average of 25% of sections from adenoviral FHIT-treated mice showed nearly normal epithelium without any obvious pathological findings; an average of 56% of sections from adenoassociated FHIT-treated mice were nearly normal; and an average of 40% of sections from the combined treatment group showed nearly normal epithelium.³⁸ These data demonstrated that Fhit expression prevents or delays NMBA-induced tumor development in Fhit + / - mice.

Recently we performed another experiment to assess the effect of FHIT gene treatment (C.M.C., unpublished data, August 2001). Fhit + / - mice received 6 doses of NMBA, as in prior experiments; 2 days after the final NMBA administration, 1 group of 14 mice received 1 dose of adenoviral FHIT virus (10¹¹ plaque-forming units/mL, 100 mL) and the second group (n = 10) received 1 dose of adenoviral green fluorescent protein virus (10¹¹ plaque-forming units/mL, 100 mL). The control group contained 13 mice that did not receive any virus. All mice were euthanized for analysis 10 weeks after virus administration. All 13 mice (100%) exposed to NMBA without virus administration had tumors and all 10 (100%) adenoviral green fluorescent protein virus–infected mice developed tumors, whereas 8 (57%) of 14 adenoviral FHIT-treated mice had tumors in the forestomach (adenoviral FHIT vs NMBA, P = .02; adenoviral FHIT vs adenoviral green fluorescent protein, P = .02). In the SCJ, 9 (69%) of 13 mice exposed to NMBA had tumors and 6 (60%) of 10 adenoviral green fluorescent protein–treated mice formed tumors, whereas 2 (14%) of 14 adenoviral FHIT-treated mice had tumors (adenoviral FHIT vs NMBA, P = .006; adenoviral FHIT vs adenoviral green fluorescent protein, P = .03). The data show that adenoviral FHIT, but not adenoviral green fluorescent protein, significantly inhibits NMBA-induced tumor in Fhit knockout mice, suggesting that FHIT prevents carcinogen-induced cancer development.

COMMENT

Jump to Section  • Top  • Introduction  • Methods  • Results  • Comment  • Author information  • References

Concerning human trials, viral vectors would be good candidates at the present time to reach high levels of gene transduction efficiency in various tissues. Adenoviral vector can provide high levels of expression in almost all tissues regardless of cell proliferation, but it can generate host immunoresponse, resulting in difficulty in repeating therapy. Adenoassociated viral vector can provide high efficiency in transducing endoluminal cells in the intestine without troublesome immunoresponse in vivo^82-83 and the expression reaches a plateau after some delay when administered to mice in gavage.³⁸ In clinical trials, a targeting approach with an endoscope, for example, may maximize efficiency and minimize undesirable responses. Although it is difficult in small animals to target lesions, endoscopes may be effective in treating carcinogen-induced cancer in humans. Preclinical studies with larger animals than mice would be appropriate to evaluate the efficacy of the targeting approach.

Effects on cell growth inhibition and apoptosis induced by FHIT seem different from those of tumor suppressor genes TP53 and RB.³⁵ Adenoviral FHIT expression was detected at 24 hours after infection, but the growth inhibition and apoptosis induction were apparently observed first at 1 to 3 days and peaked at 2 to 5 days after transduction.^35-37 Growth inhibition and apoptosis induced by exogenous TP53 expression became apparent at 24 hours and peaked at 72 hours after transduction.^84-85 The cellular responses to exogenous FHIT expression are slower than those to exogenous TP53 expression, although the magnitude of the final effect is at a similar level.³⁵ Adenoviral FHIT-transduced cells accumulate in the S phase in lung, head and neck, and some esophageal cancer cells,^35-36 whereas p53-expressing cells and Rb1-positive cells responding to effectors such as p16 expression undergo G₀-G₁ arrest.^6-7,35 The results suggest that an apoptotic pathway mediated by the FHIT gene is different from pathways mediated by tumor suppressor genes TP53 and RB. The study showed that adenoviral FHIT is an equally effective growth inhibitor in both TP53-deleted H1299 and wild-type TP53-containing A549 lung cancer cells.³⁵ It is likely that gene therapy combining both TP53 and FHIT genes would be more effective because different apoptotic pathways would be activated.³⁵

In conclusion, the FHIT gene is inactivated in a number of cancers, from premalignant to advanced stages. Many studies have shown that Fhit expression inhibits or eliminates cancer cell growth in vitro and in vivo, at least in part through caspase-dependent apoptosis. Fhit expression appears to prevent carcinogen-induced cancer development in Fhit + / - mice, in a model involving premalignant or early-stage tumor; the observation suggests that FHIT gene therapy could be a novel clinical approach in treatment of early-stage cancer and in cancer prevention. Thus, we suggest a rationale for further investigation of FHIT gene delivery as a preventive treatment as well as therapeutic strategy for carcinogen-related precancerous lesions, such as those of esophagus and lung.^{38, 86}

AUTHOR INFORMATION

Jump to Section  • Top  • Introduction  • Methods  • Results  • Comment  • Author information  • References

Author Contributions: Study concept and design: Ishii, Vecchione, Fong, Baffa, Huebner, Croce.

Acquisition of data: Ishii, Dumon, Vecchione, Fong, Baffa, Huebner, Croce.

Analysis and interpretation of data: Ishii, Dumon, Vecchione, Fong, Huebner, Croce.

Drafting of the manuscript: Ishii, Dumon, Vecchione, Fong, Croce.

Critical revision of the manuscript for important intellectual content: Ishii, Dumon, Vecchione, Fong, Baffa, Huebner, Croce.

Statistical expertise: Ishii, Dumon, Vecchione.

Obtained funding: Ishii, Dumon, Croce.

Administrative, technical, or material support: Ishii, Dumon, Fong.

Study supervision: Ishii, Dumon, Baffa, Croce.

Funding/Support: This work was supported partially by US Public Health Service grants CA39860, CA77738, CA56036, and CA83698 from the National Cancer Institute.

Financial Disclosures: Dr Croce served as a consultant for and owned stock in Idun Pharmaceuticals. Drs Huebner and Croce owned a patent for fragile histidine triad (FHIT) nucleic acids and methods of producing FHIT proteins.

Corresponding Author and Reprints: Carlo M. Croce, MD, Kimmel Cancer Center, Thomas Jefferson University, 233 S 10th St, Philadelphia, PA 19107-5799 (e-mail: C_CROCE{at}lac.jci.tju.edu).

Author Affiliations: Kimmel Cancer Institute and Kimmel Cancer Center, Thomas Jefferson University, Philadelphia, Pa.

REFERENCES

Jump to Section  • Top  • Introduction  • Methods  • Results  • Comment  • Author information  • References

1. Nowell PC, Croce CM. Chromosomes, genes, and cancer. Am J Pathol. 1986;125:7-15. ISI | PUBMED 2. Knudson AG. Antioncogenes and human cancer. Proc Natl Acad Sci U S A. 1993;90:10914-10921. FREE FULL TEXT 3. Nowell PC. Chromosomes and cancer. Adv Cancer Res. 1993;62:1-17. ISI | PUBMED 4. Roth JA, Cristiano RJ. Gene therapy for cancer. J Natl Cancer Inst. 1997;89:21-39. FREE FULL TEXT 5. Anklesaria P. Gene therapy: a molecular approach to cancer treatment. Curr Opin Mol Ther. 2000;2:426-432. ISI | PUBMED 6. Sherr CJ. Cancer cell cycles. Science. 1996;274:1672-1677. FREE FULL TEXT 7. Levine AJ. P53, the cellular gatekeeper for growth and division. Cell. 1997;88:323-331. FULL TEXT | ISI | PUBMED 8. Rocco JW, Sidransky D. p16(MTS-1/CDKN2/INK4a) in cancer progression. Exp Cell Res. 2001;264:42-55. FULL TEXT | ISI | PUBMED 9. Glover TW, Berger C, Coyle J, Echo B. DNA polymerase alpha inhibition by aphidicolin induces gaps and breaks at common fragile sites in human chromosomes. Hum Genet. 1984;67:136-142. FULL TEXT | ISI | PUBMED 10. Paradee W, Wilke CM, Wang L, et al. A 350-kb cosmid contig in 3p14.2 that crosses the t(3;8) hereditary renal cell carcinoma translocation breakpoint and 17 aphidicolin-induced FRA3B breakpoints. Genomics. 1996;35:87-93. FULL TEXT | ISI | PUBMED 11. Wilke CM, Hall BK, Hoge A, et al. FRA3B extends over a broad region and contains a spontaneous HPV16 integration site. Hum Mol Genet. 1996;5:187-195. FREE FULL TEXT 12. Rassool FV, Le Beau MM, Shen ML, et al. Direct cloning of DNA sequences from the common fragile site region at chromosome band 3p14.2. Genomics. 1996;35:109-117. FULL TEXT | ISI | PUBMED 13. Ohta M, Inoue H, Cotticelli MG, et al. The FHIT gene, spanning the chromosome 3p14.2 fragile site and renal carcinoma-associated t(3;8) breakpoint, is abnormal in digestive tract cancers. Cell. 1996;84:587-597. FULL TEXT | ISI | PUBMED 14. Strachan T, Read AP. Organization of the human genome. In: Strachan T, Read AP, eds. Human Molecular Genetics. 2nd ed. New York, NY: Wiley-Liss; 1999:139-168. 15. Huebner K, Garrison PN, Barnes LD, Croce CM. The role of the FHIT/FRA3B locus in cancer. Annu Rev Genet. 1998;32:7-31. FULL TEXT | ISI | PUBMED 16. Sozzi G, Huebner K, Croce CM. FHIT in human cancer. Adv Cancer Res. 1998;74:141-166. ISI | PUBMED 17. Sozzi G, Pastorino U, Moiraghi L, et al. Loss of FHIT function in lung cancer and preinvasive bronchial lesions. Cancer Res. 1998;58:5032-5037. FREE FULL TEXT 18. Birrer MJ, Hendricks D, Farley J, et al. Abnormal Fhit expression in malignant and premalignant lesions of the cervix. Cancer Res. 1999;59:5270-5274. FREE FULL TEXT 19. Mori M, Mimori K, Shiraishi T, et al. Altered expression of Fhit in carcinoma and precarcinomatous lesions of the esophagus. Cancer Res. 2000;60:1177-1182. FREE FULL TEXT 20. Gatalica Z, Lele SM, Rampy BA, Norris BA. The expression of Fhit protein is related inversely to disease progression in patients with breast carcinoma. Cancer. 2000;88:1378-1383. FULL TEXT | ISI | PUBMED 21. Tanimoto K, Hayashi S, Tsuchiya E, et al. Abnormalities of the FHIT gene in human oral carcinogenesis. Br J Cancer. 2000;82:838-843. FULL TEXT |