Key words: thyroid carcinomas; chromosome rearrangements; fusion transforming genes.

SUMMARY
Well differentiated thyroid carcinomas, provide a unique model of human, epithelial cell carcinogenesis. Their molecular characterization has allowed to associate specific genetic alterations to the two papillary and follicular hystotypes which, despite their common origin, display different biological and clinical behaviors. A common mechanism of oncogenic activation has been observed in these tumor, based on the peculiar characteristic of thyroid epithelium to generate fusion transforming genes by chromosomal rearrangements. The reasons for this peculiar uniqueness of thyrocytes are not known but a structural explanation, based on the spatial contiguity in the interphase nuclei of thyrocytes of the two fused genes and enzymatic features of these cells which render them apoptosis resistant to DNA damage, have been proposed to account for this behavior.

1. INTRODUCTION
Two types of well differentiated carcinomas derive from the thyroid epithelium, the papillary and the follicular carcinoma, respectively. In spite of their common origin, these tumors are regarded as different biological entities: the follicular carcinoma is associated with endemic goiter, a diet with low iodine intake and metastatizes almost exclusively via the blood stream often to bones. The papillary carcinoma, conversely, is associated with a previous radiation exposure, high iodine intake and metastatizes through lymphatic spread to regional lymph nodes.

On this ground, years ago, we and others started cytogenetic and molecular analyses aimed at the identification of the molecular mechanisms which could drive the transformation of the thyrocytes toward a papillary or a follicular phenotype.

The obtained experimental results well support the concept that different genetic lesions could determine the outcome of one or the other type of thyroid well differentiated carcinoma. Remarkably, most of the identified oncogenic alterations belong to the family of the fusion transforming genes which are generated by chromosomal rearrangements, both intrachromosomal (i.e. inversions) and interchromosomal (i.e. translocations) [1; 2].

It is noteworthy that chromosomal alterations generating fusion oncoproteins are common in leukemia/lymphomas and sarcomas but, in human carcinomas, they are almost exclusively found in thyroid well differentiated carcinomas [3]. Here the genetic alterations identified in these tumors, in particular the diverse types of fusion transforming genes generated by chromosomal rearrangements, will be described. Moreover, the peculiar susceptibility of thyroid epithelium to chromosomal alterations will be discussed.

2. FUSION TRANSFORMING GENES IN PAPILLARY THYROID CARCINOMAS.
Over the last fifteen years our laboratory and others have reported the alternative involvement of the RET and NTRK1 genes, coding for neurotrophin tyrosine kinase (TK) receptors, in the development of a consistent fraction (about 50%) of papillary thyroid carcinomas (PTCs). Somatic rearrangements, both intra and interchromosomal, of RET and NTRK1, on chromosome 10 and 1, respectively, generate several forms of fusion transforming genes. Their transforming activity was directly assessed by transfection analysis of the tumor DNA from patients with PTCs, by using NIH-3T3 recipient cells. The corresponding normal DNAs did not transform the latter [4]. In each case, the intracellular domain of the two genes, which displays a tyrosine kinase enzymatic activity, is fused to the N-terminus of different genes. We have defined the latter as "activating" genes (Fig. 1).

The RET/PTC1 oncogene represents the first example of oncogene activation in solid tumors due to an acquired chromosomal abnormality. RET/PTC1 is a chimeric transforming sequence generated by the fusion of the TK domain of RET to the 5' terminal sequence of the gene H4/D10S170. The latter has been shown to display a coiled-coil sequence, which confers to the oncoprotein the ability to form dimers, resulting in a constitutive activation of the TK function. Both partners in the fusion have been localized to chromosome 10q and their fusion is the molecular event consequent to a paracentromeric inversion of chromosome 10q, inv 10 (q11.2 q21.2) [5; 6].

We subsequently found a second example of RET activation: the RET/PTC2 oncogene. In this case the rearrangement involved the gene of the regulatory subunit RI-a of Protein Kinase A, which maps to chromosome 17q23. Interestingly, like the H4 gene, RI-a also contains a dimerization domain and the construction of RET/PTC2 mutants with deletions in RI-a, has demonstrated that the formation of dimers is necessary to express the activity of the oncogene. The cytogenetic analysis of one case of RET/PTC2 positive carcinoma revealed that this oncogene arises from a t(10;17)(q11.2;q23) reciprocal translocation [1].

Finally, a third example of RET activation in PTCs has been reported, RET/PTC3. In this oncogene, the TK domain of RET is fused to sequences derived from a previously unknown gene named ELE1 (otherwise named RFG). Interestingly, we have localized ELE1 in the same chromosomal region of RET, 10q11.2 Also in this case, a paracentric inversion of the long arm of chromosome 10 was identified [1]. A summary of the different combinations of RET with these activating genes is reported in Fig. 2. More recently, additional versions of RET rearranged with different alterated genes have been described, in particular in papillary thyroid carcinomas associated with radiation exposure (i.e. Chernobyl nuclear accident) [7]. Interestingly, the described RET rearrangements seem to be restricted to those belonging to the papillary type [8].

In our analysis of PTCs, several cases showed an activation of the NTRK1 proto-oncogene [4]. In three specimens we identified a chimeric sequence generated by the rearrangement of an isoform of non-muscle tropomyosin (TPM3) and NTRK1. The former has been mapped to chromosome 1q22-23. Therefore, the NTRK1 localization on 1q22 suggested that a 1q intrachromosomal rearrangement could have generated the TRK oncogene and further molecular analysis of TRK positive PTCs indicated that an intrachromosomal inversion, inv(1q), provided the mechanism of the NTRK1 oncogenic activation in these tumors.

In the remaining cases genes different from TPM provided the 5' terminus of the oncogene; we therefore designated the latter as TRK-T (Fig. 2). Three cases showed the fusion of NTRK1 TK domain to sequences of the TPR (Translocated Promoter Region) gene, originally identified as part of the MET oncogene. We have localized the TPR locus on chromosome 1q25. Therefore, as for TRK, an intrachromosomal rearrangement, molecularly defined as an inversion of 1q, is responsible for the formation of this fusion oncogene designated as TRK-T1. A rearrangement involving the same two genes, has been found in two other papillary thyroid tumors. Although the two rearrangements involve different genomic regions of the partner genes, they occur in the same intron of both TPR and NTRK1. As a consequence, the same mRNA and 1323 aminoacid oncoprotein are produced and designated TRK-T2 in both cases.

Finally, an additional oncogene derived from NTRK1 activation, designated TRK-T3, has been described. Sequence analysis revealed that TRK-T3 contains 1412 nucleotides of NTRK1 preceded by 598 nucleotides belonging to a novel gene named TFG (TRK Fused Gene) encoding a 68 kDa cytoplasmic protein. The latter displays, in the TFG part, a coiled-coil region that endows the oncoprotein with the capability to form complexes, as shown by the analysis of deletion and point mutants. The TFG gene is ubiquitously expressed and is located on chromosome 3 thus suggesting that a still undetected t(1q;3) occurred in that tumor [1].

The relative frequency of RET and NTRK1 activation have been found different in PTCs collected from various geographical areas. We have demonstrated the formation of oncogene sequences from RET and NTRK1 in about 50% of PTCs collected at the National Cancer Institute in Milan (Italy) with RET positive cases accounting for 34% of the total. However, lower percentages of RET activation have been described by other groups, ranging from 2.5% in Saudi Arabia to 15% in USA [1]. A possible explanation for these results could reside in exposure to diverse environmental factors. The primary role of environmental factors has already been shown in tumors of children exposed to high levels of radiation following the Chernobyl disaster and in thyroid lesions associated to radiation therapy of the head and neck. Although uncertainties remain about the dose-response relationship, history of radiation exposure, particularly in childhood, is probably the best characterized risk factor for thyroid cancer and for RET oncogenic rearrangements as well. In fact, molecular studies from both our laboratory and Klugbauers group have found RET rearrangements in 66% of post-Chernobyl PTCs. Interestingly, in these cases RET/PTC3 was the most frequent observed rearrangement [7; 9].

A comparative analysis of the oncogenes originated by the rearrangement of the two receptor tyrosine kinase proto-oncogenes RET and NTRK1, allows us to identify a common cytogenetic and molecular mechanism for their activation.

In all cases, chromosomal rearrangements mostly intrachromosal (e.g. inversion) (Fig. 1) fuse the TK portion of the RTKs to the 5' end of different genes that we have designated as "activating" genes (Fig. 2). Furthermore, although functionally different, the various activating genes share three properties, most likely related to the ttransforming activity of the fusion proteins: a) they are ubiquitously expressed, b) in all the cases, demonstrated or predicted domains able to form dimers or multimers, have been identified and, finally, c) the TK associated enzymatic activity is moved from the membrane to the cytoplasm.

The oncogenic activation of RET and NTRK1 proto-oncogenes, following chromosomal rearrangements in papillary thyroid carcinomas, can be defined therefore as an ectopic, constitutive and topologically abnormal expression of their enzymatic (TK) activity. Although radiation exposure seems definitively associated to these genetic events, the possibility that other carcinogenic factors may produce the same genetic changes remains to be determined. It seems likely that specific, as yet undetermined environmental factors are important in the etiology of RET/NTRK1 positive PTCs.

3. FUSION TRANSFORMING GENES IN FOLLICULAR THYROID CARCINOMAS.
The follicular thyroid tumors have been first molecularly characterized by identifying the presence of mutated alleles of the three members of the RAS oncogene family. In follicular carcinogenesis, RAS mutation appears to occur early, being detectable in adenomas and even in multinodular goiters and also in the last stages of thyroid tumor growth, being found in a percentage of undifferentiated thyroid carcinomas. Interestingly, the frequency of RAS activation seems to be higher in iodine-deficient area. Moreover, the presence of an activated RAS significantly correlates with the metastatic behavior, in particular bone metastasis, of follicular thyroid tumors [1].

More recently, Kroll et al [2], reported the molecular characterization of the already known t(2;3) (q13;p25) chromosomal translocation in follicular tumors [10], by showing the occurrence of a fusion product between PAX8, which encodes a paired domain transcription factor essential for thyroid development, and the peroxisome proliferator-activated receptor gamma (PPARg) genes. The fusion RNA and protein were detected in 5 out of 8 follicular carcinomas and the involvement of PPARg gene with another unidentified partner was suggested in other 2 cases. This genetic alteration was not detected in 20 follicular adenomas, 10 papillary carcinomas and 10 multinodular hyperplasias. It thus appears that the described fusion product is tumor specific and, like RET and NTRK1 rearrangements segregate with the papillary type of thyroid carcinomas, PPARg rearrangements are linked to the follicular ones. The functional analysis of the fusion product, revealed that PAX8-PPARgamma functions as a dominant negative suppressor of wild -type PPARg activities. One of the latter consists in the capability of PPARg gene product to inhibit growth and to promote differentiation, following ligand binding. It is therefore conceivable that the abrogation of normal PPARg function is important in cancer development and could play a critical role in thyroid follicular oncogenesis. If a relationship does exist between PPARg inactivation and the previously reported RAS genes mutations has not been yet investigated.

4. THYROID EPITHELIUM AND CHROMOSOMAL REARRANGEMENTS
From the results reported in the previous chapters, it can be derived that the different biochemical alterations caused by the two oncogenic processes involving tyrosine kinase activity (RET/NTRK1 activation) or gene transcription activation (PPARg activation) may account, at least in part, for the different morphology and clinical behavior displayed by papillary and follicular carcinomas (Fig. 3).

However, an important question, related to these findings, remains open: what are the unique characteristics of thyrocytes that set them apart from other epithelial cells, making them susceptible to the formation of fusion oncoproteins, following chromosomal rearrangement, a feature usually associated with mesenchymal and hematopoietic cells?

A definitive answer to this question has not been yet provided, however the previously reported association between papillary carcinomas and ionizing radiation exposure can provide some suggestions.

First of all, the analysis of the thyroid tumors following Chernobyl nuclear accident has shown a high prevalence of RET rearrangements (62.3%) with a significant predominance of RET/PTC3 over RET/PTC1 form. NTRK1 rearrangements were rare (3.3%) and, in the same percentage of cases, novel types of RET/PTCs were identified. More important, RET/PTC3 rearrangements are related to rapid tumor development and, consequently are more frequently associated with the first detected tumor cases and with a higher radiation-exposure. Overall, a genotype/phenotype evaluation of these tumors reveals a characteristic spectrum of gene rearrangements associated with typical biological and clinical tumor phenotypes [7].

A second line of evidences comes from several experimental models. After an early report showing the induction of RET/PTC1 oncogene in "in vitro" X-ray treated tumor cell lines not showing before the treatment RET activation [11], more recently, Mizuno et al reported the formation of RET/PTCs3 rearrangement in human normal thyroid tissue transplanted in scid mice after X-ray exposure of 50 Gy.

The more frequent form observed over the entire observation period (up to 60 days) was RET/PTC1, whereas RET/PTC3 was detected only 7 days after irradiation although at a lower frequency [12]. Interestingly, the same authors reported few year before that in a similar model (human normal thyroid tissue transplanted in mammary fat pad of scid mice and irradiated at a dose of 50 GY), they detected also the formation of BCR-ABL product, a fusion transforming gene associated to the pathogenetic mechanism(s) originating chronic myelogenic leukemia (CML), two days after radiation exposure but this rearranged gene was undetectable subsequently. On the contrary, RET/PTC1 was identified at day 2 and throughout the two months periods [13]. These experiments indicate that ionizing radiation causes various oncogenic activations but cells with only specific genetic alterations are specifically and selectively retained in the target tissue for which they are involved in the carcinogenic process.

Finally, two experimental models seem to provide key information for the understanding of the peculiar features of the thyroid epithelium when chromosomal rearrangements are considered. The first one approached the problem of the thyroid specificity of RET rearrangements from a structural point of view. Nikiforova et al. [14] asked, in fact, whether despite the great linear distance between them, RET and H4 recombination (forming RET/PTC1 oncogene) might be promoted by their proximity in the nucleus. By two-color FISH analysis and three-dimensional microscopy, they mapped the positions of the two genes within interphase nuclei of different cell types. They found that at least one pair of RET and H4 was juxtaposed in 35% of normal human thyroid cells, in 21% of normal lymphocytes and in 6% of normal mammary epithelial cells. Spatial contiguity of RET and H4 in thyrocytes may thus provide a structural basis for the thyroid specific generating of RET/PTC1 rearrangement by allowing a single radiation track to produce a double-strand break in each gene at the same site in the nucleus. The second model challenged the same problem by a complementary enzymatically based approach. Yang et al. [15] showed that "in vitro" exposure of human thyroid cells to ionizing radiation (up to 5-8 Gy) failed to induce apoptosis. The same result was obtained when the thyroid gland was irradiated in intact rat thus showing that thyrocytes are relatively resistant to apoptosis following DNA damage. Importantly, in both cases they found, instead, a significant increase of the DNA end-joining enzymatic activity. These results may thus reflect the occurrence of specific responses in thyroid cells following irradiation.

5. CONCLUSIONS
Although derived from the same precursor thyrocytes, papillary and follicular carcinomas display different phenotypes associated to specific genetic alterations most of which are represented by fusion transforming genes generated by different types of chromosomal rearrangements. Why thyroid epithelium is almost unique among the other epithelia for his susceptibility to respond to DNA damage with chromosomal aberrations is not completely understood.

However, both structural and enzymatic reasons have been proposed as peculiar features of thyrocyte to account for their behavior. In particular, it is suggested that following DNA damage (e.g. irradiation) instead of triggering an apoptosis pathway, thyrocytes activate DNA repair mechanisms. It is then likely that during DNA repair, inappropriate joining of potentially oncogenic DNA sequences may occur. It would thus appear that contrary to the behavior of other epithelial cells, following DNA damage, the thyrocytes solve the dilemma opting for life with chromosomal rearrangements rather than death by apoptosis.

ACKNOWLEDGEMENTS
The work of the author is partially supported by grant AIRC (Associazione Italiana Ricerca Cancro). The author thanks Mr. Mario Azzini for the preparation of the figures and Miss. Cristina Mazzadi for secretarial help.

REFERENCES
[1] M.A.Pierotti, I.Bongarzone, M.G. Borrello, A. Greco, S. Pilotti, and G. Sozzi, Cytogenetics and molecular genetics of the carcinomas arising from the thyroid epithelial follicular cells. Genes Chrom. Cancer, 16 (1996) 1-14.
[2] T.G. Kroll, P. Sarraf, C.J. Chen, E. Mueller, B.M. Spiegelman, and J.A. Fletcher, PAX8-PPARgamma1 fusion oncogene in human thyroid carcinoma. Science, 289 (2000) 1357-1360.
[3] T.H. Rabbitts, Chromosomal translocations in human cancer. Nature, 372 (1994) 143-149.
[4] I. Bongarzone, M.A. Pierotti, N. Monzini, P. Mondellini, G. Manenti, R. Donghi, S. Pilotti, M. Grieco, M. Santoro, A. Fusco, G. Vecchio, and G. Della Porta, High frequency of activation of tyrosine kinase oncogenes in human papillary thyroid carcinoma. Oncogene, 4 (1989) 1457-1462.
[5] M. Grieco, M. Santoro, M.T. Berlingieri, R.M. Melillo, R. Donghi, I. Bongarzone, M.A. Pierotti, G. Della Porta, A. Fusco, and G. Vecchio, PTC is a novel rearranged form of the ret proto-oncogene and is frequently detected in vivo in human thyroid papillary carcinomas. Cell, 60 (1990) 557-563.
[6] M.A. Pierotti, M. Santoro, R.B. Jenkins, G. Sozzi, I. Bongarzone, M. Grieco, N. Monzini, M. Miozzo, M.A. Herrmann, A. Fusco, I.D. Hay, G. Della Porta, and G. Vecchio, Characterization of an inversion on the long arm of chromosome 10 juxtaposing D10S170 and ret and creating the oncogenic sequence ret/ptc. Proc. Natl. Acad. Sci. USA, 89 (1992) 1616-1620.
[7] H.M. Rabes, E.P. Demidchik, J.D. Sidorow, E. Lengfelder, C. Beimfohr, D. Hoelzel, and S. Klugbauer, Pattern of radiation-induced RET and NTRK1 rearrangements in 191 post-Chernobyl papillary thyroid carcinomas: biological, phenotypic, and clinical implications. Clin. Cancer Res., 6 (2000) 1093-1103.
[8] M. Santoro, F. Carlomagno, I.D. Hay, M.A. Herrmann, M. Grieco, R. Melillo, M.A. Pierotti, I. Bongarzone, G. Della Porta, N. Berger, J.L. Peix, C. Paulin, N. Fabien, G. Vecchio, R.B. Jenkins, and A. Fusco, Ret oncogene activation in human thyroid neoplasms is restricted to the papillary cancer subtype. J. Clin. Invest., 89 (1992) 1517-1522.
[9] L. Fugazzola, S. Pilotti, A. Pinchera, T.V. Vorontsova, P. Mondellini, I. Bongarzone, A. Greco, L. Astakhova, M.G. Butti, E.P. Demidchik, F. Pacini, and M.A. Pierotti, Oncogenic rearrangements of the RET proto-oncogene in papillary thyroid carcinomas from children exposed to Chernobyl nuclear accident. Cancer Res., 55 (1995) 5617-5620.
[10] G. Sozzi, M. Miozzo, C.T. Cariani, I. Bongarzone, S. Pilotti, M.A. Pierotti, and G. Della Porta, A t(2;3)(q12-13;p24-25) in follicular thyroid adenomas. Cancer Genet. Cytogenet., 63 (1992) 38-41.
[11] T. Ito, T. Seyama, K.S. Iwamoto, T. Hayashi, T.,Mizuno, N. Tsuyama, K. Dohi, N. Nakamura, and M. Akiyama, In vitro irradiation is able to cause RET oncogene rearrangement. Cancer Res., 53 (1993) 2940-2943.
[12] T. Mizuno, K.S. Iwamoto, S. Kyoizumi, H. Nagamura, T. Shinohara, K. Koyama, T. Seyama, and K. Hamatani, Preferential induction of RET/PTC1 rearrangement by X-ray irradiation. Oncogene, 19 (2000) 438-443.
[13] T. Mizuno, S. Kyoizumi, T. Suzuki, K.S. Iwamoto, and T. Seyama, Continued expression of a tissue specific activated oncogene in the early steps of radiation-induced human thyroid carcinogenesis. Oncogene, 15 (1997) 1455-1460. [14] M.N. Nikiforova, J.R. Stringer, R. Blough, M. Medvedovic, J.A. Fagin, and Y.E. Nikiforov, Proximity of chromosomal loci that participate in radiation-induced rearrangements in human cells. Science, 290 (2000) 138-141.
[15] T.T. Yang, H. Namba, T. Hara, N. Takmura, Y. Nagayama, S. Fukata, N. Ishikawa, K. Kuma, K. Ito, and S. Yamashita, p53 induced by ionizing radiation mediates DNA end-jointing activity, but not apoptosis of thyroid cells. Oncogene, 14 (1997) 1511-1519.

Figure 1

Figure 2

Figure 3

Back . . .