RET Rearrangements in Radiation-Induced Papillary Thyroid Carcinomas: High Prevalence of Topoisomerase I Sites at Breakpoints and Microhomology-Mediated End Joining in ELE1 and RET Chimeric Genes

Genomics 73, 149 –160 (2001) doi:10.1006/geno.2000.6434, available online at http://www.idealibrary.com on

RET Rearrangements in Radiation-Induced Papillary Thyroid Carcinomas: High Prevalence of Topoisomerase I Sites at Breakpoints and Microhomology-Mediated End Joining in ELE1 and RET Chimeric Genes S. Klugbauer,* ,1 P. Pfeiffer,† ,1 H. Gassenhuber,‡ C. Beimfohr,* and H. M. Rabes* ,2 *Institute of Pathology, Ludwig Maximilians University of Munich, Munich, Germany; †Institute of Cell Biology, Medical School of the University of Essen, Essen, Germany; and ‡Munich Infomation Center for Protein Sequences, Munich, Germany Received August 21, 2000; accepted November 3, 2000

Children exposed to radioactive iodine after the Chernobyl reactor accident frequently developed papillary thyroid carcinomas (PTC). The predominant molecular lesions in these tumors are rearrangements of the RET receptor tyrosine kinase gene. Various types of RET rearrangements have been described. More than 90% of PTC with RET rearrangement exhibit a PTC1 or PTC3 type of rearrangement with an inversion of the H4 or ELE1 gene, respectively, on chromosome 10. To obtain closer insight into the mechanisms underlying PTC3 inversions, we analyzed the genomic breakpoints of 22 reciprocal and 4 nonreciprocal ELE1 and RET rearrangements in 26 post-Chernobyl tumor samples. In contrast to previous assumptions, an accumulation of breakpoints at the two Alu elements in the ELE1 sequence was not observed. Instead, breakpoints are distributed in the affected introns of both genes without significant clustering. When compared to the corresponding wildtype sequences, the majority of breakpoints (92%) do not contain larger deletions or insertions. Most remarkably, at least one topoisomerase I site was found exactly at or in close vicinity to all breakpoints, indicating a potential role for this enzyme in the formation of DNA strand breaks and/or ELE1 and RET inversions. The presence of short regions of sequence homology (microhomologies) and short direct and inverted repeats at the majority of breakpoints furthermore indicates a nonhomologous DNA end-joining mechanism in the formation of chimeric ELE1/Ret and Ret/ELE1 genes. © 2001 Academic Press

INTRODUCTION

Children exposed to radioactive fallout after the Chernobyl reactor accident frequently developed pap1

These authors contributed equally to this work. To whom correspondence and reprint requests should be addressed at Institute of Pathology, Ludwig Maximilians University of Munich, Thalkirchner Strasse 36, D-80337 Munich, Germany. Telephone: (49) 89-5160-4081. Fax: (49) 89-5160-4083. E-mail: [email protected]. 2

illary thyroid carcinomas (PTC), in particular in the highly contaminated regions of Belarus (Pacini et al., 1997; Rabes et al., 2000). Molecular analysis of postChernobyl PTC revealed a high prevalence of RET rearrangements (Rabes and Klugbauer, 1997, 1998; Nikiforov et al., 1997; Smida et al., 1999). NTRK1 rearrangements and other aberrations known to be involved in thyroid tumorigenesis are rare or absent (Rabes and Klugbauer, 1998; Beimfohr et al., 1999; Rabes, in press). Proto-RET encodes a membrane-associated receptor tyrosine kinase (TK) that is expressed at specific developmental stages in subsets of neural crest-derived cells (Avantaggiato et al., 1994). In oncogenic rearrangements, the amino-terminal domain is deleted and the remaining TK domain is fused to the 5⬘-end of various genes that are ubiquitously expressed and contain dimerization domains. Thus, the RET rearrangement leads to an expression of chimeric mRNAs and proteins in thyroid epithelial cells, a cell type not derived from the neural crest. The fusion proteins exhibit intrinsic and constitutive TK activity, proven to be sufficient to induce PTC in transgenic mice (Saggartz et al., 1997; Powell et al., 1998). The precise mechanisms of generating chromosomal RET rearrangements are still obscure. One reason might be that only the final rearranged state is accessible to analysis while the exact site and structure of the initial lesion remain unknown. However, it is generally accepted that processes catalyzed by the cellular DNA repair and recombination machinery are involved in these mechanisms. The high prevalence of RET rearrangements in PTC of children exposed to radioiodine after the Chernobyl reactor accident supports the notion that this radionuclide induces DNA doublestrand breaks (DSB), which may be considered the primary cause of these rearrangements. According to current models, the repair of DSB is achieved by at least three major pathways: (i) homologous recombination, (ii) single-strand annealing (SSA), and (iii) nonhomologous DNA end joining (NHEJ) (Kanaar et al.,

149

0888-7543/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.

150

KLUGBAUER ET AL.

1998; Pfeiffer et al., 2000). While the former two depend on extensive regions of sequence homology and the genes of the Saccharomyces cerevisiae Rad52 epistasis group (Haber, 1999; van Dyck et al., 1999), NHEJ, the major pathway of DSB repair in mammalian cells, does not require extensive sequence homologies and is able to rejoin broken DNA ends directly. NHEJ appears to comprise at least two different pathways: one that is accurate and involves the Ku70/80 heterodimer and another that is Ku-independent and error-prone (Critchlow and Jackson, 1998; Go¨ttlich et al., 1998; Pfeiffer, 1998; Feldmann et al., 2000). The error-prone pathway forms mainly short deletions. The resulting breakpoints often coincide with short patches of sequence homology (2–9 bp), also called “microhomologies,” which are also present in the former wildtype sequences (Roth and Wilson, 1986; Thacker et al., 1992; Nicola´s et al., 1995; Go¨ttlich et al., 1998; Feldmann et al., 2000). Since these features indicate a direct involvement of microhomologies (or direct repeat units) in junction formation (Pfeiffer et al., 2000), this pathway was also designated “direct repeat end joining” (DREJ) (Thacker, 1999) or microhomology-mediated SSA (Go¨ttlich et al., 1998; Feldmann et al., 2000). So far, we have investigated 191 PTC of children and adolescents exposed to radioactive fallout up to an age of 18 years at the time of the Chernobyl reactor accident and found RET rearrangements in 49.2% of the tumors (Rabes et al., 2000). In the meantime, six different types of RET rearrangements have been reported for childhood thyroid carcinomas; 91.5% of RET rearrangement-positive tumors contained a PTC1 or PTC3 rearrangement involving the genes for H4 and RET or ELE1 and RET, respectively. Four variants of the PTC3 rearrangement have been observed, designated PTC3r1, r2, r3, and PTC4 (for review Rabes and Klugbauer, 1998). To gain closer insights in the mechanisms underlying the formation of the PTC3 type of RET rearrangement, we have analyzed the genomic sequences at the gene fusion points of 26 childhood tumors of Belarus. Our data provide evidence that DREJ (Thacker, 1999) or microhomology-mediated SSA (Go¨ttlich et al., 1998; Feldmann et al., 2000) is most likely responsible for the formation of the majority of the PTC3 rearrangements investigated here. MATERIALS AND METHODS Thyroid tumors. Papillary thyroid carcinomas were obtained from patients who lived in highly contaminated areas of Belarus after the Chernobyl reactor accident. They underwent thyroidectomy at the Department of Surgery, Medical High School of Minsk. The following patients were included: M2, male, 28 years old at the time of the reactor explosion, thyroidectomized on 12/4/1992; M12, male, 10 years, 4/19/1993; M36, female, 1.5 years, 6/18/1993; M80, male, 2 years, 3/20/1995; M81, male, 3.5 years, 3/22/1995; M89, male, 2 years, 7/5/1995; M122, female, 6 years, 11/8/1995; M124, male, 0.75 years, 11/13/1995; M129, male, 3 years, 12/1/1995; M145, female, 4.5 years, 1/29/1996; M147, female, 6 years, 1/30/1996; M151, female, 3.5 years, 2/13/1996; M153, male, 10.5 years, 2/19/1996; M160, female, 5.5 years, 3/25/1996; M161, female, 16 years, 3/29/1996; M162, female, 0.5 years, 4/1/1996; M190, female, 12 years, 7/24/1996; M214,

female, 0.2 years, 11/8/1996; M216, female, 3 years, 11/11/1996; M219, male, 2.5 years, 11/18/1996; M224, male, 15 years, 11/25/1996; M225, female, 15 years, 11/27/1996; M259, male, 1 year, 2/12/1997; M263, female, 3 years, 2/19/1997; M285, male, 19.5 years, 4/18/1997; M309, male, 2 years, 6/11/1997. Purification of genomic DNA, PCR analysis, and sequencing. To identify the sequences of ELE1/RET and RET/ELE1 fusion points, genomic DNA was isolated according to described procedures (Klugbauer et al., 1996, 1998) from 26 tumor samples that were shown earlier by RT-PCR to harbor a PTC3 rearrangement (Klugbauer et al., 1995; Rabes et al., 2000). The genomic regions of ELE1/RET, RET/ELE1, and ELE1 were amplified by PCR. A thermostable DNA polymerase mixture (Roche Diagnostics GmbH, Mannheim, Germany) and, depending on the PTC3 variant, the following exonspecific primer pairs were used: rfg2/retc3 (PTC3r1,3) and PTC3a/ retc3 (PTC3r2), tm1/rfg7R (PTC3r1,3), and tm1/rfg2R (PTC3r2) (Klugbauer et al., 1995, 1998). The PCR, optimized for fidelity and yield [initial denaturation at 94°C for 10 min followed by 40 cycles (94°C for 30 s, 60°C for 30 s, 68°C for 3 min)], was performed in a GeneAmp 2400 thermal cycler (Applied Biosystems, Weiterstadt, Germany). The resulting PCR fragments, between 158 and 3384 bp in length, were sequenced directly using amplification and internal primers spaced across the relevant regions of ELE1 and RET genomic sequences. Sequence data of these parts of ELE1 and RET for the synthesis of internal primers were derived from the EMBL Data Library [Accession Nos. X77859 and X77860 (Klugbauer et al., 1996)]. Further primer sequences and annealing positions are available upon request. Computer-assisted sequence analysis. Computer-assisted services of the German Cancer Research Center Heidelberg and of the MIPS Institute in Martinsried were used. The sequences were further analyzed using the programs of the GCG package (GCG, Madison, WI).

RESULTS

PCR and Sequencing Analysis In all variants of the PTC3 type of RET rearrangements investigated so far (summarized in Rabes and Klugbauer, 1998; Rabes, in press), RT-PCR analysis has revealed breakpoints in or between three exons of the ELE1 gene, the precise genomic structure of which is unknown, and exons 11 and 12 of the RET gene (Fig. 1). Most tumor samples contained two reciprocal transcripts designated ELE1/RET and RET/ELE1: the first consisting of the 5⬘-part of ELE1 fused to the 3⬘-part of RET (exons 12–20) and the second consisting of the 5⬘-part of RET (exons 1–11) fused to the 3⬘-part of ELE1. The finding of two reciprocal transcripts together with the fact that ELE1 and RET are both located on the long arm of chromosome 10 (10q11.2) at a minimum distance of about 500 kb (Minoletti et al., 1994) suggests that a balanced intrachromosomal inversion comparable to that described for PTC1 (Pierotti et al., 1992) is responsible for the PTC3 rearrangement. At the genomic level, the DNA regions of interest span about 2.3 kb of the ELE1 gene and 2.1 kb of the RET gene. The positions of the breakpoints of the 26 PTC3 rearrangements analyzed in this study together with those of 15 previously described post-Chernobyl tumors [CH4, CH8, CH10 (Bongarzone et al., 1997) and C2, C8, C10, C11, C14, C15, C17, C20, C24, C27, C28, C30 (Nikiforov et al., 1999)] and some “spontaneous”

FIG. 1. Breakpoint distribution in PTC3 rearrangements of 26 post-Chernobyl PTC. The sites of the ELE1/RET and RET/ELE1 breakpoints are marked in the corresponding ELE1 and RET genomic regions. Designation of the individual PTC at the breakpoint position (vertical lines). Nonreciprocal cases are underlined. The breakpoint sites of six spontaneous PTC are also included (C96, C108, C117, C132, C161, and C211) (Bongarzone et al., 1997) (in red) together with those of three earlier cases (CH4, CH8, and CH10) (Bongarzone et al., 1997) (in red) and 12 cases (C2, C8, C10, C11, C24, C15, C17, C20, C24, C27, C28, and C30) (Nikiforov et al., 1999) (in blue) of post-Chernobyl PTC published recently (Bongarzone et al., 1997; Nikiforov et al., 1999).

BREAKPOINT ANALYSIS OF RADIATION-INDUCED PTC3 REARRANGEMENTS

151

152

KLUGBAUER ET AL.

FIG. 2. Sequences of the 22 reciprocal PTC3 inversions. Sixteen cases of PTC3r1 variants involve a rearrangement in the intron L (iL) of ELE1 and exon 11 (e11) or intron 11 (i11) of RET. Five cases of PTC3r2 variants involve intron K (iK) or exon L (eL) of ELE1 and intron 11 (i11) of RET. One case of a PTC3r3 variant involves eL of ELE and i11 of RET (Klugbauer et al., 1998). Sequences are given as the oncogenic ELE1/RET fusion (top sequence) and the reciprocal RET/ELE1 fusion (bottom sequence) with the ELE1 sequence part highlighted in boldface letters. Vertical lines with arrowheads (black for ELE1/RET, white for RET/ELE1) and small numbers in italics indicate the exact nucleotide position of the breakpoints in the corresponding wildtype sequences. Common sequence motifs between reciprocal fusion points are highlighted by shaded boxes. (Note that the spacing between the microhomology patches is not always identical in the two breakpoint sequences. In some cases, alternative motifs (open boxes) are possible, e.g., M36T.) Topo I sites are underlined (strong: boldface continuous lines; weak or intermediate: broken lines). Note that in some fusions, only parts of the original Topo I site are maintained (e.g., M129T). Below each fusion sequence, the alignments of the corresponding wildtype sequences are given with microhomology patches highlighted by white letters in black boxes. Note that only one alignment is shown, although in some cases at least two alternative alignments are possible. Horizontal arrows indicate direct (dr) or inverted repeats (ir) of at least 4 bp in length. Numerals on the left side of the alignment indicate the position of the left-most nucleotide in the corresponding wildtype ELE or RET sequence. M36 and M160T represent more complex rearrangements: in M36, the second breakpoint in the ELE1 sequence is not shown because it is located 84 bp further downstream; the deletion of part of the ELE eL sequence in ELE1/RET of M160T may have arisen by loop formation as shown.

rearrangements [patients without radiation history; C96, C108, C117, C132, C161, C211 (Bongarzone et al., 1997)] are presented in Fig. 1. It is obvious that the breakpoint sites are spread over a large part of the

genomic DNA without evidence for localization at exactly the same position or significant clustering. In 22 of 26 tumor samples, we obtained both ELE1/ RET and RET/ELE1 PCR fragments from genomic

BREAKPOINT ANALYSIS OF RADIATION-INDUCED PTC3 REARRANGEMENTS

153

FIG. 2—Continued

DNA and cDNA (Figs. 1 and 2; reciprocal rearrangements). In the remaining four samples (M147T, M285T, M309T, and M124T), only one PCR product was detected using genomic DNA as a template (Figs. 1 and 2; nonreciprocal rearrangements). RT-PCR yielded both cDNA fragments in three of these four samples (M147T, M285T, and M309T). The fact that one of the corresponding genomic fragments in the three tumor samples could not be amplified is possibly due to modification of PCR-primer binding sites. A single sample (M124T) did not allow amplification of the RET/ELE1 cDNA or of the corresponding genomic PCR fragment, indicating the presence of a larger deletion or an unbalanced structural rearrangement. With the exception of two more complex rearrangements (M36T and M160T) (Klugbauer et al., 1998), the

sequence alterations detected in the 26 tumor samples comprise only minor deletions or insertions, which are summarized in Table 1. In the ELE1 genomic sequence, 14 of 20 reciprocal rearrangements (70%) displayed no modification at all or small deletions (2–25 bp), while insertions of 1–5 bp were present in 6 samples (30%). In the 2.1-kb RET region, 14 of 20 reciprocal rearrangements (70%) showed no modifications at all or small deletions (1– 8 bp), and 6 (30%) exhibited small insertions (2– 6 bp). Breakpoint Analysis A computer-assisted analysis of the 2.3-kb ELE1 genomic sequence using the Husar program package revealed a low G⫹C content, especially in the corre-

154

KLUGBAUER ET AL.

FIG. 2—Continued

sponding introns K and L (Fig. 1, 35 and 39%). The larger intron (iL) contains two Alu elements, which we found to have the highest identity to the human Alu-Sb subfamily (EMBL Data Library, Accession No. gi551537). By contrast, the 2.1-kb RET genomic sequence displays a high G⫹C content in intron 11 (61%) and exons 11 (62%) and 12 (51%) (Fig. 1). The differences in G⫹C content indicate that there are no extensive regions of continuous sequence homology between ELE1 and RET. This finding is confirmed by direct sequence comparison of the corresponding wildtype sequences using the Husar DotPlot program, which shows that they do not share continuous sequence homologies longer than 9 bp (data not shown). Figures 2 and 3 show the breakpoint sequences of all 22 reciprocal rearrangements and the 4 nonreciprocal

rearrangements, respectively, and the alignments of the corresponding wildtype ELE1 and Ret sequences. The schemes reveal several remarkable phenomena. First, 20 of the 22 reciprocal cases show that identical sequence motifs consisting of 2–10 bp of uninterrupted homology were generated at both breakpoints (see shaded boxes in Fig. 2). These motifs are even more extended if small gaps of 1–2 bp are allowed. In 15 cases, both breakpoints are located directly within or adjacent to such a motif (2-bp motif, M122T; 3-bp motif, M89T, M225T, and M259T; 4-bp motif, M2T, M151T, M153T, M216T, and M219T; 5-bp motif, M190T; 6-bp motif, M12T and M161T; 7-bp motif, M145T and M224T; 10-bp motif, M160T). In 5 cases, only one breakpoint coincides with a common motif while the other is located 1–2 bp away (2-bp motif,

BREAKPOINT ANALYSIS OF RADIATION-INDUCED PTC3 REARRANGEMENTS

155

FIG. 2—Continued

M81T and M263T; 3-bp motif, M80T, M129T, and M124T). In only 2 cases, both breakpoints lie outside of the common motif (M162T and M36). Second, sequence alignments of the corresponding wildtype sequences show that all breakpoints are located directly within or between small patches of sequence homology shared by both genes (Figs. 2 and 3; see black boxes with white letters). Taking into account that the two fusion points of each reciprocal rearrangement involve four breakpoints in the corresponding wildtype sequences, the 22 reciprocal rearrangements involve a total of 88 breakpoints. The expected probability of a homology of x nucleotides to occur by chance at a breakpoint in a DNA duplex of unbiased sequence composition is given by the equation P(x) ⫽ (x ⫹ 1)(1/ 4) x(3/4) 2 [(x ⫹ 1) is the number of different ways that chance identities could yield the specified homology, (1/4) x is the probability that x nucleotides match, and (3/4) 2 is the probability that nucleotides flanking the matching nucleotides do not match (Roth et al., 1985)]. This allows one to calculate the numbers of the 88 breakpoints expected to be located within a given microhomology and compare them with the observed numbers as derived from the alignments shown in Fig. 2: 0 bp (no homology): 49 exp vs 26 obs; 1 bp: 25 exp vs 6 obs; 2 bp: 10 exp vs 14 obs; 3 bp: 3 exp vs 18 obs; 4 bp: 1 exp vs 20 obs; 5 bp: ⬍0 exp vs 4 obs. Interestingly, the observed numbers of breakpoints that coincide with a microhomology in-

crease with increasing size of the microhomology, which is inversely proportional to the expected values. This result strongly indicates that microhomologies are important for the process of rearrangement formation between Ret and ELE1. Third, using the Husar repeat stem loop program, we found that almost every breakpoint is located at or near direct and/or inverted repeat units that occur in one or the other wildtype sequence (see horizontal arrows in the alignments of Figs. 2 and 3). In most cases, parts of the sequence motifs found in the direct and inverted repeats of one wildtype sequence can be used for the formation of microhomology patches with the other wildtype sequence [e.g., Fig. 2, M2T: the direct repeat unit CAGGG, which occurs three times in Ret in close vicinity of the breakpoints, is homologous to CAGG in ELE1, and the inverted repeat GGTC-(x) 7GACC, which occurs once in ELE1, is homologous to GGTC-(x) 5-GAC in Ret]. This finding indicates that direct and inverted repeats contained in one wildtype sequence may increase the chance that microhomology patches form between the two genes. An additional screen for enzyme recognition sites using the Husar FindPatterns program revealed the presence of at least one topoisomerase I (Topo I) consensus sequence (Been et al., 1984) at or near breakpoints in all 26 samples (Table 2 and see boldface and dashed underlining in Figs. 2 and 3). Depending on the

156

KLUGBAUER ET AL.

TABLE 1 Sequence Modifications at Breakpoints Compared to the Wildtype Genomic Sequences ELE1

RET

Tumor

Modification

Sequence

Position

Modification

M2T M12T M36T M80T M81T M89T M122T M129T M145T M151T M153T M160T

No modification Deletion Deletion Deletion Deletion Deletion Deletion Deletion Insertion Insertion Insertion Deletion Insertion Insertion Deletion Insertion Deletion Deletion Deletion Deletion Insertion Deletion Deletion

— CTTC 84 bp TGCCTC TCCCT TTTTTTTT CCAATTA TTTTTTTTTTTTT AA CAG CCTA CAGACCTTGGAGAACAGT GGTCTGTT GCCAG TTTTTTTTC A CTTATA GGCAA CAAAAAAGAAAAAGTGCAGACAGAA CTCTTGGAAG GG GCTTA CA

— 659–662 41–123 913–918 661–665 790–795 485–491 772–792 395–396 33–35 670–673 ex124–141 ex144–int7 1013–1017 342–350 1658 43–48 1640–1644 1589–1616 405–414 1506–1507 516–565 361–363

Deletion No modification Deletion Deletion Insertion No modification Deletion Deletion Insertion Deletion No modification Deletion

M161T M162T M190T M214T M216T M219T M224T M225T M259T M263T

Deletion Deletion No modification Insertion Deletion Deletion Insertion Insertion Insertion Insertion

Sequence GG — GTGG GAG TCT — CAGGT TGCCCAG CG GG — AGATGACA A CC — ATC CCCTGT TGAGC ATC CAGG CT TCTGGA

Position 1565–1566 — 168–171 954–956 1048–1050 — 1068–1070 884–890 505–506 595–596 — 1714–1721 943 413–414 — 1013–1015 1786–1791 839–843 1013–1015 586–589 30–31 464–469

Note. Sequence modifications were not determined in tumor samples M124T, M147T, M285T, and M309T.

preference of the enzyme to cut at a particular sequence, three types of consensus sequences (strong, intermediate, and weak) have been described (Been et al., 1984). Interestingly, most of the breakpoints exhibit strong (13 of 26) or intermediate (6 of 26) Topo I sites, indicating that DNA cleavage by this enzyme might play a role in the formation of the PTC3 rear-

FIG. 3.

rangements investigated here. In contrast, we found only one Topo II site (Spitzner and Muller, 1988) located in the 2.3-kb ELE1 region near the breakpoint of one sample (M151T; exon l position 53–70, 17 bp downstream). This indicates that Topo II probably does not play any role in the formation of rearrangements in the series of tumors investigated here, although this en-

Fusion sequences of the four nonreciprocal PTC3 rearrangements. For explanation of symbols, see legend to Fig. 2.

BREAKPOINT ANALYSIS OF RADIATION-INDUCED PTC3 REARRANGEMENTS

TABLE 2 Different Types of Topo I Consensus Sequences at or near Breakpoints of ELE1 and RET Genomic Sequence Topo I consensus

Tumor samples

a

M89T, M122T, M129T, M153T, M190T, M216T, M219T, M263T, M36T, M214T, M224T, M160T, M309T M80T, M81T, M124T, M161T, M162T, M145T M2T, M12T, M285T M151T M147T, M225T, M259T

b

c d e

Note. For exact positions, see Figs. 2 and 3. Topo I sites are shown in generalized form [(A, T)/(C, G)/(A, T)/T], which must be read as ACAT or AGTT, etc. a Strong Topo I consensus sequence at breakpoint (A, T)/(C, G)/(A, T)/T. b Intermediate Topo I consensus sequence at breakpoint (A, T)/(C, G, A)/(A, T)/(T, C). c Weak Topo I consensus sequence at breakpoint (A, T, G)/(C, G, A)/(A, T)/(T, C). d Strong Topo I consensus sequence near breakpoint. e Weak Topo I consensus sequence near breakpoint.

zyme induces DSB (Wang, 1985) and is therefore considered a likely candidate to be involved in nonhomologous DNA end joining. DISCUSSION

A detailed analysis of ELE1/RET breakpoints at the genomic DNA level of 26 radiation-induced PTC that developed in Belarus after the Chernobyl reactor accident revealed significant features: (i) Compared to wildtype sequences, the rearrangements contain no or minor sequence alterations with only a few nucleotides lost or inserted at the breakpoints. This finding confirms previous reports (Smanik et al., 1995; Bongarzone et al., 1997; Klugbauer et al., 1998; Nikiforov et al., 1999). (ii) It is shown that most breakpoints occur within or in close vicinity of microhomology patches present in the wildtype sequences. These often comprise small direct and/or inverted repeats. (iii) As an additional new finding, we describe that one or both breakpoints of 85% of the investigated PTC3 rearrangements coincide with at least one Topo I site; the breakpoints of the remaining 15% are located in close vicinity to at least one Topo I site. When reevaluating the previously published PTC1 and PTC3 breakpoints, we found that all are associated with at least one Topo I site (Smanik et al., 1995; Bongarzone et al., 1997; Nikiforov et al., 1999). In a previous study (Bongarzone et al., 1997), breakpoints of three previously described post-Chernobyl PTC samples (CH4, CH8, and CH10) together with those of six spontaneous rearrangements were investigated. The authors suggested a significant accumulation of breakpoints in or close to the two Alu elements in the involved ELE1 intron, especially in the postChernobyl tumors. Similar results were published re-

157

cently (Nikiforov et al., 1999). In addition, these authors (Nikiforov et al., 1999) claimed a stable pattern of corresponding breakpoints by alignment of the participating RET and ELE1 introns in the opposite direction. However, neither a specific involvement of Alu elements nor any constant topographic intra-intronic relation of corresponding breakpoints was evident in the breakpoint distribution pattern in our study (Fig. 1). We did not recognize a specific clustering of breakpoints at any region in the 2.3-kb ELE1 fragment. None of the seven breakpoints found in the two Alu elements is located in the 26-bp Alu core consensus sequence (Deininger et al., 1981; Rudiger et al., 1995). In the 2.1-kb RET genomic sequence, breakpoints are also unevenly distributed (Fig. 1). The apparently random distribution of breakpoints is confirmed by computer-assisted sequence analysis, which did not reveal clustering at specific sequence motifs (e.g., A/T-rich sequences, Alu elements, consensus sequences for enzymes or transcription factors). Moreover, the comparative homology analysis between the two gene fragments did not show regions of extensive sequence homology (⬎9 bp). Therefore it is unlikely that the rearrangements observed in our samples involve Alu elements directly or are the result of homologous recombination, which is assumed to require at least 40 – 100 bp of identical sequence (Baumann and West, 1998; Benson et al., 1998). The most obvious inductor for DNA damage in postChernobyl thyroid tumors is ionizing radiation (IR) as a consequence of incorporation of radioactive iodine into the thyroid gland. Of the wide spectrum of lesions caused by IR, so-called multiply damaged sites are assumed to be biogically most important because they may ultimately lead to DSB (Ward, 1988). In addition, the significant abundance of strong and intermediate Topo I cleavage sites at the breakpoints of the PTC3 rearrangements points to still another mechanism leading to DNA strand breaks. DNA topoisomerases are ubiquitous enzymes that change the superhelical state of DNA during replication, recombination, chromosome segregation, and transcription (Bullock et al., 1985). Topo I usually generates a reversible single-strand break (SSB) immediately next to the 3⬘-end of their recognition sequence by formation of a covalent bond between the 3⬘-phosphate of the cleaved strand and a tyrosine residue of the enzyme (Been and Champoux, 1984). Upon covalent cross-linking of the enzyme by an ionizing event during S-phase, an unsealed SSB may lead to replication arrest and can be transformed into a DSB (Thacker and Ganesh, 1990; Kollmannsberger et al., 1999). Interestingly, Topo I has been reported to cause chromosomal aberrations by initiating illegitimate recombination (Bullock et al., 1985; Zhu and Schiestl, 1996; Pourquier et al., 1997). This would confirm our notion that Topo I cleavage may be involved in the formation of the PTC3 rearrangements investigated here. This possibility could be tested in human thyroid xenograft mouse model systems now available (Mizuno et al., 2000).

158

KLUGBAUER ET AL.

FIG. 4. Model for the formation of reciprocal RET/ELE1 inversions by end joining at sites of microhomologies. Close spatial proximity of both ELE and RET (green and blue arrows, respectively) within the same higher-order chromatin loop may facilitate the rearrangement. One or more ionizing events (flash) occur at the cross-over point of the loop and induce one (left side) or two independent DSB (right side). Alternatively, DSB can arise at Topo I sites during replication (see text). One DSB (step 1: here in the RET gene, blue lines) would be sufficient to initiate duplex invasion in an intact duplex (step 2: here in the ELE1 gene, green lines) at regions of sequence homology (red; arrowheads point in the 3⬘-direction) to create an inversion (step 3). Two DSB occurring simultaneously in ELE1 and RET generate four ends (step 1) that can anneal inversely with one another at short direct repeats (step 2) to create an inversion (step 3).

Chromatin is arranged in various higher-order looped structures of 2–5 Mb that consist of 30 –50 smaller 60- to 90-kb loops and are associated with transcription “factories” (Iborra et al., 1997). In addition to transcriptionally active enzymes, these factories may contain other important DNA processing enzymes such as helicases and topoisomerases, capable of cutting and facilitating exchanges between DNA molecules (Bryant, 1998b). Because both are located at the

same chromosome band (10q11.2) at a minimum distance of about 500 kb (Minoletti et al., 1994), it can be envisaged that ELE1 and RET are temporarily arranged close to each other within the same loop domain (Fig. 4). A DSB occurring within the loop may serve as a signal for the simultaneous activation of several repair and recombination pathways that may generate chromosomal aberrations by rejoining originally unrelated DSB (Bryant, 1998a; Pfeiffer et al., 2000). It is

BREAKPOINT ANALYSIS OF RADIATION-INDUCED PTC3 REARRANGEMENTS

important to note that a single DSB suffices to induce a PTC3 rearrangement although inversions are usually considered to be the result of two independent DSB (Bryant, 1998b). Based on these findings, the following initiating events may be envisaged (Fig.4, step 1): (i) A single DSB is induced by an ionizing event in the genomic sequence of the ELE1 or RET gene, leading to a signal that triggers a second DSB as a consequence of the recombinational process. (ii) It is also possible, however, that two independent DSB, one in ELE1 and the other in RET, are generated (e.g., by IR and/or Topo I). The presence of the majority of breakpoints at microhomology patches, direct repeats and inverted repeats indicates that the PTC3 inversions can arise by microhomology-driven SSA (Nicola´s et al., 1995; Go¨ttlich et al., 1998; Thacker, 1999; Feldmann et al., 2000). In this process, patches of microhomology are thought to be exposed on DNA single strands generated by exonucleolytic resection or helicase-mediated duplex unwinding. Subsequent pairing at sites of microhomology may be achieved either by direct annealing of two originally unrelated single strands (resection annealing after two initiating DSB) or by invasion of single strands into an intact, but unrelated DNA duplex (after a single initiating DSB) (Fig. 4, steps 2–3). In summary, our description of structural similiarities in a larger number of tumors induced under comparable conditions provides a new basis for an improved functional analysis of the formation of radiation-induced gene rearrangements. ACKNOWLEDGMENTS This work was supported by grants (to H.M.R.) from Deutsche Krebshilfe, Bonn, Germany, and a fellowship (to P.P.) from the Heisenberg Program of the Deutsche Forschungsgemeinschaft, Bonn, Germany. We are grateful to Professor E. Lengfelder and Professor E. Demidchik for providing samples of papillary thyroid carcinomas and to Professor P. Virsik-Peuckert and Professor D. Harder for helpful discussions on radiation-induced DNA repair. The excellent technical assistance of Andrea Eberl, Dr. Sibylle Liebmann, and Michael Ruiter is gratefully acknowledged. Thanks are due to the Otto Hug-Strahleninstitut and Christine Frenzel for support of this work.

REFERENCES Avantaggiato, V., Dathan, N. A., Grieco, M., Fabien, N., Lazzaro, D., Fusco, A., Simeone, A., and Santoro, M. (1994). Developmental expression of the RET proto-oncogene. Cell Growth Differ. 5: 305– 311. Baumann, P., and West, S. C. (1998). Role of the human Rad51 protein in homologous recombination of double-strand break repair. Trends Biochem. Sci. 23: 247–251. Been, M. D., and Champoux, J. J. (1984). Breakage of singlestranded DNA by eukaryotic type 1 topoisomerase occurs only at regions with the potential for base-pairing. J. Mol. Biol. 180: 515–531. Been, M. D., Burgess, R. R., and Champoux, J. J. (1984). Nucleotide sequence preference at rat liver and wheat germ type 1 DNA topoisomerase breakage sites in duplex SV40 DNA. Nucleic Acids Res. 12: 3097–3114.

159

Beimfohr, C., Klugbauer, S., Demidchik, E. P., Lengfelder, E., and Rabes, H. M. (1999). NTRK1 re-arrangement in papillary thyroid carcinomas of children after the Chernobyl reactor accident. Int. J. Cancer 80: 842– 847. Benson, F. E., Baumann, P., and West, S. C. (1998). Synergistic actions of Rad51 and Rad52 in recombination and DNA repair. Nature 391: 401– 404. Bongarzone, I., Butti, M. G., Fugazzola, L., Pacini, F., Pinchera, A., Vorontsova, T. V., Demidchik, E. P., and Pierotti, M. A. (1997). Comparison of the breakpoint regions of ELE1 and RET genes involved in the generation of RET/PTC3 oncogene in sporadic and in radiation-associated papillary thyroid carcinomas. Genomics 42: 252–259. Bryant, P. E. (1998a). Mechanisms of radiation-induced chromatid breaks. Mutat. Res. 404: 107–111. Bryant, P. E. (1998b). The signal model: A possible explanation for the conversion of DNA double-strand breaks into chromatid breaks. Int. J. Radiat. Biol. 73: 243–251. Bullock, P., Champoux, J. J., and Botchan, M. (1985). Association of crossover points with topoisomerase I cleavage sites: A model for nonhomologous recombination. Science 230: 954 –958. Critchlow, S. E., and Jackson, S. P. (1998). DNA end-joining: From yeast to man. Trends Biochem. Sci. 23: 394 –398. Deininger, P. L., Jolly, D. J., Rubin, C. M., Friedmann, T., and Schmid, C. W. (1981). Base sequence studies of 300 nucleotide renatured repeated human DNA clones. J. Mol. Biol. 151: 17–33. Feldmann, E., Schmiemann, V., Goedecke, W., Reichenberger, S., and Pfeiffer, P. (2000). DNA double-strand break repair in cell-free extracts from Ku80-deficient cells: Implications for Ku serving as an alignment factor in nonhomologous DNA end joining. Nucleic Acids Res. 28: 2585–2596. Go¨ttlich, B., Reichenberger, S., Feldmann, E., and Pfeiffer, P. (1998). Rejoining of DNA double-strand breaks in vitro by single-strand annealing. Eur. J. Biochem. 258: 387–395. Haber, J. E. (1999). DNA repair. Gatekeepers of recombination. Nature 389: 665– 667. Iborra, F. J., Pombo, A., Jackson, D. A., and Cook, P. R. (1997). Active RNA polymerases are localized within discrete transcription “factories” in human nuclei. J. Cell Sci. 109: 1427–1436. Kanaar, R., Hoeijmakers, J. H. J., and van Gent, D. C. (1998). Molecular mechanisms of DNA double strand break repair. Trends Cell Biol. 8: 483– 489. Klugbauer, S., Lengfelder, E., Demidchik, E. P., and Rabes, H. M. (1995). High prevalence of RET rearrangement in thyroid tumors of children from Belarus after the Chernobyl reactor accident. Oncogene 11: 2459 –2467. Klugbauer, S., Lengfelder, E., Demidchik, E. P., and Rabes, H. M. (1996). A new form of RET rearrangement in thyroid carcinomas of children after the Chernobyl reactor accident. Oncogene 13: 1099 – 1102. Klugbauer, S., Demidchik, E. P., Lengfelder, E., and Rabes, H. M. (1998). Molecular analysis of new subtypes of ELE/RET rearrangements, their reciprocal transcripts and breakpoints in papillary thyroid carcinoma of children after Chernobyl. Oncogene 16: 671– 675. Kollmannsberger, C., Mross, K., Jakob, A., Kanz, L., and Bokemeyer, C. (1999). Topotecan—A novel topoisomerase I inhibitor: Pharmacology and clinical experience. Oncology 56: 1–12. Minoletti, F., Butti, M. G., Coronelli, S., Miozzo, M., Sozzi, G., Pilotti, S., Tunnacliffe, A., Pierotti, M. A., and Bongarzone, I. (1994). The two genes generating RET/PTC3 are localized in chromosomal band 10q11.2. Genes Chromosomes Cancer 11: 51–57. Mizuno, T., Iwamoto, K. S., Kyoizumi, S., Nagamura, H., Shinohara, T., Koyama, K. Seyama, T., and Hamatani, K. (2000). Preferential induction of RET/PTC1 rearrangement by X-ray irradiation. Oncogene 19: 438 – 443. Nicola´s, A. L., Munz, P. L., and Young, C. S. H. (1995). A modified single-strand annealing model best explains the joining of DNA

160

KLUGBAUER ET AL.

double-strand breaks in mammalian cells and cell extracts. Nucleic Acids Res. 23: 1036 –1043. Nikiforov, Y. E., Rowland, J. M., Bove, K. E., Monforte-Munoz, H., and Fagin, J. A. (1997). Distinct pattern of ret oncogene rearrangements in morphological variants of radiation-induced and sporadic thyroid papillary carcinomas in children. Cancer Res. 57: 1690 – 1694. Nikiforov, Y. E., Koshoffer, A., Nikiforova, M., Stringer, J., and Fagin, J. A. (1999). Chromosomal breakpoint positions suggest a direct role for radiation in inducing illegitimate recombination between the ELE1 and RET genes in radiation-induced thyroid carcinomas. Oncogene 18: 6330 – 6334. Pacini, F., Vorontsova, T., Demidchik, E. P., Molinaro, E., Agate, L., Romei, C., Shavrova, E., Cherstvoy, E. D., Ivashkevitch, Y., Kuchinskaya, E., Schlumberger, M., Ronga, G., Filesi, M., and Pinchera, A. (1997). Post-Chernobyl thyroid carcinoma in Belarus children and adolescents: Comparison with naturally occurring thyroid carcinoma in Italy and France. J. Clin. Endocrinol. Metab. 82: 3563–3569. Pfeiffer, P. (1998). The mutagenic potential of DNA double-strand break repair. Toxicol. Lett. 96/97: 119 –129. Pfeiffer, P., Goedecke, W., and Obe, G. (2000). Mechanisms of DNA double-strand break repair and their potential to induce chromosomal aberrations. Mutagenesis 15: 289 –302. Pierotti, M. A., Santoro, M., Jenkins, R. B., Sozzi, G., Bongarzone, I., Grieco, M., Monzini, N., Miozzo, M., Herrmann, M. A., Fusco, A., Hay, I. D., Della Porta, G., and Vecchio, G. (1992). 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: 1616 –1620. Pourquier, P., Pilon, A. A., Kohlhagen, G., Mazumder, A., Sharma, A., and Pommier, Y. (1997). Trapping of mammalian topoisomerase I and recombinations induced by damaged DNA containing nicks or gaps. Importance of DNA end phosphorylation and camptothecin effects. J. Biol. Chem. 272: 26441–26447. Powell, D. J., Jr., Russell, J., Nibu, K., Li, G., Rhee, E., Liao, M., Goldstein, M., Keane, W. M., Santoro, M., Fusco, A., and Rothstein, J. L. (1998). The RET/PTC3 oncogene: Metastatic solid-type papillary carcinomas in murine thyroids. Cancer Res. 58: 5523– 5528. Rabes, H. M. (in press). Gene rearrangements in radiation-induced thyroid carcinomas. Med. Pediatr. Oncol., in press. Rabes, H. M., and Klugbauer, S. (1997). Radiation-induced thyroid carcinomas in children: High prevalence of RET rearrangement. Verh. Dtsch. Ges. Path. 81: 139 –144. Rabes, H. M., and Klugbauer, S. (1998). Molecular genetics of childhood papillary thyroid carcinomas after irradiation: High prevalence of RET rearrangement. Rec. Res. Cancer Res. 154: 248 –264. Rabes, H. M., Demidchik, E. P., Sidorow, J. D., Lengfelder, E., Beimfohr, C., Hoelzel, D., and Klugbauer, S. (2000). Pattern of radiation-induced RET and NTRK1 rearrangements in 191 post-

Chernobyl papillary thyroid carcinomas: Biological, phenotypic and clinical implications. Clin. Cancer Res. 6: 1093–1103. Roth, D. B., Porter, T. N., and Wilson, J. H. (1985). Mechanisms of nonhomologous recombination in mammalian cells. Mol. Cell. Biol. 5: 2599 –2607. Roth, D. B., and Wilson, J. H. (1986). Nonhomologous recombination in mammalian cells: Role for short sequence homologies in the joining reaction. Mol. Cell. Biol. 6: 4295– 4304. Rudiger, N. S., Gregersen, N., and Kielland-Brandt, M. C. (1995). One short well conserved region of Alu-sequences is involved in human gene rearrangements and has homology with prokaryotic chi. Nucleic Acids Res. 23: 256 –260. Saggartz, J. E., Jhiang, S. M., Tong, Q., and Capen, C. C. (1997). Thyroid-stimulating hormone promotes growth of thyroid carcinomas in transgenic mice with targeted expression of the ret/PTC1 oncogene. Lab. Invest. 76: 307–318. Smanik, P. A., Furminger, T. L., Mazzaferri, E. L., and Jhiang, S. M. (1995). Breakpoint characterization of the ret/PTC oncogene in human papillary thyroid carcinoma. Hum. Mol. Genet. 4: 2313– 8. Smida, J., Salassidis, K., Hieber, L., Zitzelsberger, H., Kellerer, A. M., Demidchik, E. P., Negele, T., Spelsberg, F., Lengfelder, E., Werner, M., and Bauchinger, M. (1999). Distinct frequency of ret rearrangements in papillary thyroid carcinomas of children and adults from Belarus. Int. J. Cancer 80: 32–38. Spitzner, J. R., and Muller, M. T. (1988). A consensus sequence for cleavage by vertebrate DNA topoisomerase II. Nucleic Acids Res. 16: 5533–5556. Thacker, J. (1999). Repair of ionizing radiation damage in mammalian cells. Alternative pathways and their fidelity. C. R. Acad. Sci. 322: 103–108. Thacker, J., and Ganesh, A. N. (1990). DNA-break repair, radioresistance of DNA synthesis, and camptothecin sensitivity in the radiation-sensitive irs mutants: Comparisons to ataxia–telangiectasia cells. Mutat. Res. 235: 49 –58. Thacker, J., Chalk, J., Ganesh, A., and North, P. (1992). A mechanism for deletion formation in DNA by human cell extracts: The involvement of short sequence repeats. Nucleic Acids Res. 20: 6183– 6188. van Dyck, E., Stasiak, A. Z., Staziak, A., and West, S. C. (1999). Binding of double-strand breaks in DNA by human Rad52 protein. Nature 398: 728 –731. Wang, J. C. (1985). DNA topoisomerases. Annu. Rev. Biochem. 65: 635– 692. Ward, J. F. (1988). DNA damage produced by ionizing radiation in mammalian cells: Identities, mechanisms of formation, and reparability. Progr. Nucleic Acids Res. Mol. Biol. 35: 95–125. Zhu, J., and Schiestl, R. H. (1996). Topoisomerase I involvement in illegitimate recombination in Saccharomyces cerevisiae. Mol. Cell. Biol. 16: 1805–1812.