Oncogene database

2248

FGF3

fibroblast growth factor 3;FGF3;fibroblast growth factor 3

FGF-3|INT-2 proto-oncogene protein|V-INT2 murine mammary tumor virus integration site oncogene homolog|fibroblast growth factor 3 (murine mammary tumor virus integration site (v-int-2) oncogene homolog)|heparin-binding growth factor 3|murine mammary tumor

11q13

protein-coding

BACKGROUND: Int-2 (fibroblast growth factor-3) is a gene that belongs to the fibroblast growth factor gene family. It has been implicated in the carcinogenesis of several types of cancer, including esophageal squamous cell carcinoma, breast, head, and neck and lung carcinomas; but no firm data on its biological activity regarding neoplasms arising from the glandular epithelia of the gastrointestinal tract exists. METHODS: In the present immunohistochemical study, we investigated the presence of int-2 encoded protein in a panel of 80 cases of colon carcinoma of various stages, grades and sizes. A sheep antihuman int-2 antibody was applied to paraffin-embedded tissue sections from the tumor samples. The percentage of int-2 immunostaining in the positively stained specimens was evaluated by image analysis. RESULTS: Int-2 was positively detected in only four tumors (i.e. 5% of the cases examined). ALL immunopositive cases were moderately differentiated tumors; the adjacent mucosa did not express int-2 protein. The relevant patients were male. CONCLUSION: These findings suggest that the role of int-2 in colorectal carcinogenesis is probably a limited one.

The int-2 proto-oncogene encodes several products related to the fibroblast growth factor (FGF) family. FGFs have been associated with mesoderm induction in the amphibian embryo and int-2 has a distinct pattern of expression throughout development in vertebrates. But evidence for a function of int-2 in embryo-genesis has been lacking. In the mouse embryo, int-2 transcripts have been detected in the rhombencephalon at a developmental stage where classical experiments showed that the induction of the inner ear occurs. This raises the possibility that int-2 may constitute a signal for the induction of the otic vesicle, the primordium of the inner ear. We provide direct evidence for this view by showing that (1) the formation of the otic vesicle is inhibited by antisense oligonucleotides targeted to the secreted form of int-2, and by antibodies against int-2 oncoproteins, and (2) basic FGF (bFGF) can mimic the inductive signal in the absence of the rhombencephalon.

mutations of ras oncogenes are believed to play an important role in the initiation or progression of human tumors. In thyroid tumors the incidence of ras activation by specific point mutations has been reported to range from 33% in follicular adenomas up to 60% in anaplastic carcinomas. Because of our long-standing interest in Hurthle cell tumors, we began a study of 70 such cases to determine the incidence of ras mutations and their clinical correlates. Analysis of N-ras sequences at codon position 61, with the polymerase chain reaction method and oligonucleotide probe hybridization, showed point mutations of the normal codon CAA* in eight tumor samples. One was a mutation from CAA to AAA, one from CAA to CTA,* and six from CAA to CGA. These mutations would result in amino acid substitutions of lysine, leucine, or arginine for the normal glutamine at position 61 in the N-ras protein. Identical ras mutations in two tumors and some of their surrounding thyroid tissue may indicate that activating ras point mutations are an early event in carcinogenesis. The incidence of mutations was 1 of 24 (4%) of the histologically benign tumors, 5 of 34 (15%) of the intermediate tumors (with vascular or capsular permeation), and 2 of 12 (17%) in the malignant group. Four of these eight patients died of metastatic thyroid disease and four are alive without evidence of recurrence.

The classification of int-2 as a growth factor is based primarily on the similarities between the predicted amino acid sequence and that of basic fibroblast growth factor (bFGF), as well as other members of this expanding family of related proteins. In this review, we summarise the background to the identification of int-2 as a proto-oncogene in virally induced mouse mammary tumours and describe key features of the structure and expression of both the mouse and human homologues. The normal sites of int-2 expression include specific embryonic cell types suggesting multiple inductive or morphogenetic roles. Recent progress in the characterisation of the int-2 product will be discussed in relation to the similarities and differences between int-2 and other FGFs.

Products of the fibroblast growth factor-related proto-oncogene int-2 have been detected by using a monoclonal antibody and polyclonal antisera raised against synthetic peptides predicted from the DNA sequence. COS-1 monkey cells transfected with int-2 DNA linked to the simian virus 40 early promoter contained at least four int-2-specific proteins, presumably representing modified forms of the expected 27-kilodalton primary translation product. The level of expression was increased approximately six- to eightfold by mutation of sequences around the presumed initiation codon, negating their capacity to encode a short oligopeptide in the +1 reading frame. Both tunicamycin inhibition and in vitro translation experiments indicated that some of the modifications correspond to asparagine-linked glycosylation, for which the sequence predicts a single site. In line with the similarities between INT-2 and other fibroblast growth factors, the in vitro translation products functioned as weak mitogens for mammary epithelial cells.

The FGF-related proto-oncogene int-2 is implicated in mouse embryogenesis, since it is expressed in specific tissues during gastrulation and neurulation (Wilkinson et. al. 1988). Here, we describe the expression of this gene during subsequent fetal development, int-2 transcripts are restricted to Purkinje cells in the cerebellum and to regions of the developing retina containing early-stage differentiating cells. This high level expression is not detected in the mature cerebellum or retina. In addition, int-2 RNA is detected in the mesenchyme of the developing teeth and in sensory regions of the inner ear. This complex and dynamic pattern suggests multiple roles of this proto-oncogene during fetal development of the mouse.

Gene mapping experiments show that Fis-1, a preferred integration region for Friend murine leukemia virus in hematopoietic neoplasms, is extremely closely linked to Int-2, a preferred integration region for mouse mammary tumor virus in mammary carcinomas. Studies at the RNA and DNA level prove that these loci are distinct.

A major proportion of carcinomas induced by mouse mammary tumour virus (MMTV) show evidence for proviral activation of a cellular gene, int-2, on chromosome 7. The sequence of 7869 bp of DNA spanning the transcription unit of int-2 was determined and compared with that of a series of int-2-specific cDNA clones derived from mammary tumour RNA. The predicted positions of intron-exon boundaries, established by alignment of cDNA and chromosomal DNA sequences, indicate that the gene comprises at least three exons. An open reading frame capable of encoding a protein of 245 amino acids with an estimated mol. wt of 27 kd, is flanked by substantial non-coding segments at both 5 and 3 ends. Comparison of the chromosomal DNA sequence and the predicted amino acid sequence with available data-bases has revealed no homology to other known genes. These results are discussed in relation to the status of int-2 as a candidate proto-oncogene.

int-2 is one of two cellular genes (int-1 and int-2) currently implicated in the genesis of mammary carcinomas by mouse mammary tumor virus and may constitute a novel cellular proto-oncogene. Using low-stringency hybridization with mouse int-2 probes, we established that homologous genes exist in a variety of mammalian species, including humans, but failed to detect related sequences in other classes and phyla. Recombinant bacteriophage clones and a single cosmid encompassing the human int-2 gene were isolated and characterized by restriction enzyme mapping. A survey of nine primary human breast tumors, three breast tumor cell lines, and three normal individuals revealed no evidence for gross amplification or rearrangement of the int-2 locus. Three distinct restriction fragment length polymorphisms were observed which could prove useful in future linkage studies. By a combination of in situ hybridization of metaphase chromosomes and somatic cell genetics, the human int-2 gene was mapped to chromosome 11, band q13.

Gene targeting--homologous recombination of DNA sequences residing in the chromosome with newly introduced DNA sequences--in mouse embryo-derived stem cells promises to provide a means to generate mice of any desired genotype. We describe a positive nd negative selection procedure that enriches 2,000-fold for those cells that contain a targeted mutation. The procedure was applied to the isolation of hprt- and int-2- mutants, but it should be applicable to any gene.

The proto-oncogene int-2 has been implicated in the formation of mouse mammary-tumour-virus-induced mammary tumours. Analysis of the predicted coding sequence indicates that int-2 is a member of the fibroblast growth factor family. Previous studies using Northern blot analysis suggested that normal expression of int-2 may be confined to extra-embryonic endoderm lineages of embryonic stages of mouse development. We have used in situ hybridization and Northern blot analysis to examine directly int-2 expression in embryo stem cells and in the developing embryo from early gastrulation to midsomite stages. Complex patterns of accumulation of int-2 RNA were observed in embryonic and extra-embryonic tissues. The data suggest multiple roles for int-2 in development which may include migration of early mesoderm cells and induction of the otocyst.

A growing body of evidence demonstrates the relevant role of the int-2 (FGF-3) oncogene in human carcinomas. To investigate its angiogenic activity, the human epithelial mammary cell line MCF-10A was infected with a retroviral expression vector carrying the int-2 oncogene. Infected cells were entrapped in an alginate pellet and placed on the chorioallantoic membrane of chick embryos. After 7 days, a dense capillary network was found to grow toward the pellet, whereas parental cells did not show any angiogenic activity. Conditioned medium from int-2-infected cells was injected i.p. twice daily into rats over a period of 10 days. The mesentery of treated rats showed numerous small blood vessels originating from larger vascular arcades and growing through the stromal layer of the mesentery. In control experiments, neither medium for cell culture nor conditioned medium from parental cells was found to induce angiogenesis. In conclusion, the stimulation of blood vessel growth by int-2-infected cells suggests that the production of the int-2 protein is associated with the acquisition of the angiogenic phenotype.

The human colon carcinoma cell line, SW613-S, is composed of cells with a high-level amplification of the MYC proto-oncogene that are tumorigenic in nude mice and of cells with a low-level amplification of MYC that are not tumorigenic. Transcripts from FGF-3, a member of the fibroblast growth factor gene family, accumulate in cells from tumorigenic clones, but are undetectable in those from non-tumorigenic clones. Nuclear run-on analyses indicate that this differential FGF-3 expression is regulated at the level of transcription initiation. Determination of the structure of the FGF-3 transcripts indicates that they are generated by splicing of the three exons and termination at the single polyadenylation site predicted from the genomic sequence. Their size heterogeneity is due to multiple initiation sites spanning a 700 base-pair long promoter region. FGF-3 is activated in tumors induced in nude mice by MYC-transfected cells from non-tumorigenic clones. However, in most of the cell lines established from these tumors, FGF-3 expression tends to be lost upon in vitro propagation. Thus, in these transfectant cell lines, the presence of exogenous MYC gene copies is not sufficient to activate FGF-3 expression and in vivo growth is also required.

Transgenic mice harboring the int-2/Fgf-3 protooncogene under transcriptional control of the mouse mammary tumor virus (MMTV) promoter/enhancer exhibit a dramatic, benign hyperplasia of the mammary gland. In one int-2 transgenic line (TG.NX), this growth disturbance is evoked by pregnancy and regresses after parturition. Regression of hyperplastic mammary epithelium is less complete after successive pregnancies, and, within 10 months, most TG.NX mice stochastically develop mammary carcinomas that are transplantable in virgin, syngeneic mice. To identify genes that cooperate with int-2 in cell transformation, we infected TG.NX transgenic mice with MMTV. In a cohort of 14 animals, most mammary tumors represented clonal or oligoclonal outgrowths harboring one to five proviral MMTV integrants. Eight of 35 (23%) MMTV+ tumors exhibited proviral insertion at the Wnt-1 locus. No provirus was detected at the int-2, int-3, or Wnt-3 loci. By Southern analysis, two tumors had proviral insertions at the same genomic location, which was mapped to chromosome 15. Cloning of this int locus identified an additional member of the Wnt gene family. The predicted 389-amino acid protein is most closely related to zebrafish Wnt-10a (58% amino acid identity over 362 residues) and, based on homology analysis, was designated Wnt-10b. This newly discovered Wnt family member was expressed in the embryo and mammary gland of virgin but not pregnant mice and represents a candidate collaborating oncogene of int-2/Fgf-3 in the mammary epithelium.

We derived mice that carry a targeted insertion of a neor gene in the int-2 (Fgf-3) proto-oncogene coding sequences. The mutation was found to be recessive and mice that were homozygous for the insertion did not often survive to adulthood. The mutant mice had defects in the development of the tail and inner ear that could be correlated with disruption of int-2 expression in the posterior primitive streak and hindbrain or otic vesicle. While the tail phenotype was 100% penetrant, we found that the inner ear phenotype had reduced penetrance and variable expressivity. The variable expressivity could not be attributed to variability in the genetic background of the mutant allele or to leaky expression from the mutant allele. Thus, we conclude that even in a uniform genetic background, stochastic variation in the expression of a developmental circuit can result in dramatic differences in phenotypic consequences.

Fibroblast growth factors (FGFs), such as basic FGF, have been implicated in the growth of Kaposi s sarcoma (KS) cells in vitro. In the evaluation of the expression of the various genes of the different members of the FGF family and their receptors in fresh KS tissue specimens, int-2 was found to be expressed in more than half of the KS tumors examined. Using reverse transcription PCR, the expression of int-2 was detected in 21 of 38 (55.2%) fresh KS biopsy specimens. In contrast, int-2 mRNA transcripts were not found in normal appearing skin from the same patients except in one sample which was obtained from an AIDS patient with disseminated KS lesions. Sequence data confirmed that the amplified sequences were derived from int-2 mRNA with proper splicing. In addition, 12 nucleic acid alterations were identified in eight out of nine KS tumor samples sequenced. Using immunohistochemical methods, int-2 protein was detected in some of the spindle-shaped tumor cells surrounding the abnormal endothelial-lined vascular slits histologically characteristic of KS. Int-2 specific immunostaining was shown to be present in both the nuclei and cytoplasm of these spindle cells but was more pronounced in the nuclei. Neither amplification nor gross rearrangement of the int-2 gene was detected in KS lesions by Southern blot analysis. These results suggest that the expression of int-2 may play a role in the pathogenesis KS by stimulating local angiogenesis and cell proliferation.

Amplification of the int-2 oncogene was observed in 10 out of 78 (13%) samples of DNA isolated from cells of primary breast tumor. A clinico-morphological examination established a correlation between this molecular-genetical anomaly and tumor size: gene amplification was detected in 1 out 34 (3%) cases of breast tumor under 2 cm and in 9 out 44 (20%) tumors measuring 2 cm and more (p < 0.05). Similarly, a kind of correlation was found between gene amplification and low differentiation of breast cancer cells: enhanced copying ability of the oncogene was registered in 2 out 4 (50%) breast tumors with low cell differentiation, 8 out 69 (12%) tumors with medium cell differentiation, while it was not identified in 5 breast carcinomas with high cell differentiation (p < 0,2).

The role of int-2 oncogene amplification on the prognosis of breast cancer patients was investigated in 128 patients with node-negative primary breast cancers given first-line local-regional treatments until relapse and with a median follow-up of 65 months. tumours had been previously characterised for oestrogen (ER) and progesterone receptor (PgR) status and proliferative activity (3H-thymidine labelling index). Amplification of the int-2 oncogene occurred in 18% of cases and was significantly related to the presence of hormone receptors and to menopausal status or age, but not to proliferative status. Patients with tumours exhibiting int-2 amplification had a lower probability of disease-free survival than patients with non-amplified tumours and frequently developed local-regional recurrence. disease-free survival analysis, adjusted for the prognostic contribution provided by tumour size, steroid receptors and proliferative rate, indicated that the association between int-2 amplification and risk of relapse was maintained and remained constant even in the presence of the other co-variates. Interestingly, int-2 amplification was a further prognostic discriminant within subsets of patients with a putatively good (i.e., tumour size <20 mm, ER+ and PgR+) or poor prognosis (i.e., high labelling index). Our exploratory study suggests that within node-negative patients, int-2 amplification could be a valuable and independent prognosticator, useful to identify patients at high risk of local-regional recurrence.

Constitutive tyrosine phosphorylation of CrkL was recently demonstrated in platelets from chronic myelogenous leukaemia (CML) patients but BCR-ABL tyrosine kinase could not be detected in the platelet lysates. We studied platelets from 14 CML patients with different types of BCR-ABL mRNA and with maximal platelet counts ranging from 149 to 3069 x 10(9)/l. P210BCR-ABL protein was detected by Western blotting in platelet lysates of 12/13 CML patients with active disease but not in the lysate of platelets from a Ph-positive acute lymphoblastic leukaemia (ALL) patient in remission or eight BCR-ABL-negative controls including one essential thrombocythaemia (ET) patient. Immunoblotting of p210BCR-ABL-positive platelets lysates with anti-CrkL antibody revealed a CrkL triplet consisting of one unphosphorylated and two phosphorylated forms of the protein. This CrkL phosphorylation pattern was not observed in normal platelets or CML platelets treated with ABL tyrosine kinase inhibitor CGP57148B. The presence of BCR-ABL provides an explanation for the constitutive tyrosine phosphorylation of CrkL in CML platelets. As no correlation was observed between platelet counts and platelet BCR-ABL protein expression, thrombocytosis or thrombocythaemia in CML cannot be explained by constitutive BCR-ABL-mediated CrkL tyrosine phosphorylation.