¹Department of Biochemistry, Niigata University School of Medicine, Asahimachi 1-757, Niigata 951-8122, Japan

²Division of Life Science, Graduate School of Science and Technology, Niigata University, Ikarashi 2-8050, Niigata 950-2181, Japan

³Radiation Biology Center, Kyoto University, Yoshida-Konoecho, Sakyou-Ku 606-8315, Japan

Correspondence to: Ryo Kominami, Department of Biochemistry, Niigata University School of Medicine, Asahimachi 1-757, Niigata 951-8122, Japan

Our previous genome-wide analysis of allelic loss for thymic lymphomas that were induced by -irradiation in F₁ hybrid mice between BALB/c and MSM strains suggested the centromeric region on chromosome 11 as a site harboring a tumor suppressor gene. Interestingly, to this region the mouse Ikaros gene was mapped which was postulated to participate in oncogenic process from the study of Ikaros knockout mice. Here we show fine allelic loss mapping in the vicinity of Ikaros in 191 lymphomas, indicating that the critical region of allelic loss was centered at the Ikaros locus. PCR analysis revealed that nine lymphomas failed to give PCR-amplification for either of two exon primer pairs, indicative of homozygous deletion. Six and five mutations were detected in the N-terminal zinc finger domain and the activation domain of Ikaros, respectively, and six of the eleven were frameshift or nonsense mutations that resulted in truncation of Ikaros protein. The results strongly suggest a direct role for Ikaros in development of mouse thymic lymphomas. This provides the experimental basis for further analysis of Ikaros mutations in human cancer.

tumor suppressor gene; Ikaros; allelic loss (or LOH) analysis; somatic mutation; -ray-induced thymic lymphoma

Introduction

Identification of novel tumor suppressor genes is important for understanding the molecular mechanism underlying carcinogenesis, but only about 25 such genes have been identified. Tumor suppressor genes often map to chromosomal regions that exhibit allelic loss or loss of heterozygosity. Accordingly, mapping of allelic loss regions among multiple tumors leads to identification of new loci harboring tumor suppressor genes (Knudson, 1985; Weinberg, 1991; Vogelstein and Kinzler, 1993). Mouse genetic systems have a number of advantages for such allelic loss mapping, including the availability of thousands of polymorphic markers and the ability to breed genetically uniform and informative hybrid mice that can provide an essentially unlimited number of tumors (Kemp et al., 1993; Dietrich et al., 1994). We previously performed a genome-wide allelic loss analysis for murine thymic lymphomas that were induced by -irradiation in F₁ hybrid mice between BALB/c and MSM strains (Matsumoto et al., 1998). Among 62 microsatellite loci examined, three loci, D11Mit71, D12Mit181 and D16Mit122, exhibited high frequencies of allelic loss, 40, 65 and 45%, respectively. This suggested that novel tumor suppressor genes may reside in these chromosomal regions.

The Ikaros gene encodes, by alternative splicing, a family of zinc finger proteins essential for the development of the lymphoid system (Georgopoulos et al., 1992, 1994; Molnar and Georgopoulos, 1994). Mice homozygous for a deletion in the Ikaros DNA-binding domain fail to generate mature lymphocytes, whereas heterozygous mice possess lymphocytes with normal cell surface antigens during the first month of their lives but undergo dramatic changes in T cell populations shortly afterward (Georgopoulos et al., 1994; Winandy et al., 1995). The heterozygotes develop lympho-proliferative disorders and ultimately die of T-cell leukemias and lymphomas. Interestingly, this gene is mapped near the D11Mit71 locus at the centromeric end of mouse chromosome 11 (Molnar et al., 1996). This raised a possibility that the Ikaros gene is involved in the lymphoma development and prompted us to perform a fine mapping of both the Ikaros locus and the peak region of the allelic loss in the thymic lymphomas. Here we provide the mapping data and also demonstrate homozygous deletions and somatic mutations of the Ikaros gene in the lymphomas which indicate the function loss of Ikaros in the tumors. These results strongly suggest that Ikaros functions as a tumor suppressor gene in the development of -ray-induced mouse thymic lymphomas.

Results

Our initial genome-wide scanning of radiation-induced thymic lymphomas revealed a frequent allelic loss region at the centromeric region of mouse chromosome 11 (Matsumoto et al., 1998). To further localize the region, we generated 205 thymic lymphomas by -irradiation of F₁ hybrid mice between female BALB/c and male MSM heterozygous for the p53-deficient allele (see Materials and methods). Of these, we obtained 191 lymphomas without marked contamination of normal tissues and used for the allelic loss analysis. Among the lymphomas, 116 lymphomas were from F₁ mice carrying a p53-deficient allele and the other 75 from p53 wild-type mice. Also, three markers, D11Mit133, D11Mit63 and D11Mit227, between D11Mit71 and D11Mit19 on chromosome 11 and the Ikaros gene were mapped using 220 backcross mice between BALB/c and MSM, since these markers were clustered in a single locus in the MIT map. As for the Ikaros gene mapping, a HpaII sensitivity polymorphism on exon 4 was used. The revised order was as follows: Centromere- D11Mit71 -5.0 cM- D11Mit62 -1.8 cM- Ikaros -0.9 cM- D11Mit133 -0.9 cM- D11Mit63/D11Mit227 -11 cM- D11Mit19.

Allelic loss analysis was carried out using the five markers and the Ikaros primers. Figure 1 shows gel electrophoretic patterns of D11Mit62 and D11Mit133 PCR products and HpaII digests of the Ikaros exon 4 PCR products. Allelic difference was clearly detected by the former two polymorphic markers and also by the exon 4 PCR products which had three HpaII-recognition sites in MSM but one in BALB/c. Among 191 lymphomas, allelic loss analysis at the Ikaros locus was successfully conducted for 182 cases. The frequency of allelic loss was 99/182 (54%) at the Ikaros locus, higher than that of the other four loci. The remaining nine lymphomas failed to give clear band-signals because of homozygous deletion. Figure 2 shows a compilation of allelic losses at the Ikaros and two flanking marker loci which allows us to classify lymphomas into five types of A to E (see the legend to Figure 2). The 14 type-E lymphomas showed allelic loss only at the Ikaros locus. These results indicated that the critical region of allelic loss was centered at the Ikaros locus, bordered by D11Mit62 and D11Mit133, suggesting that Ikaros was a candidate in the development of -ray induced lymphomas.

Homozygous deletion of the Ikaros gene was examined with PCR under the condition of 30 amplification cycles for 108 lymphomas (99+9 cases) that were selected by allelic loss screening in the region surrounding the Ikaros locus. Four pairs of primers were used, two pairs being on exon 4 and exon 7 of Ikaros and the other two on flanking regions, the two ends of a BAC insert containing the Ikaros gene (Figure 3a). As a positive control, primers for the catenin gene were always included in the PCR reaction mixture. Figure 3b shows gel patterns of PCR products of six lymphomas. The first three lymphomas yielded clear bands for all four Ikaros sites as well as the catenin bands, indicative of no homozygous deletion. On the other hand, Lymphomas 4 and 6 failed to give band signals by either of the four Ikaros primer pairs and Lymphoma 5 did not yield clear bands for either of the two exon primers but was positive for the two flanking primer pairs. There were six other lymphomas devoid of signals; five were of Lymphoma 4 (or 6) type and one of Lymphoma 5 type (data not shown). Under the condition of 40 PCR amplification cycles, all of these nine lymphomas were positive for band-signals. However, these signals were due to contamination by surrounding normal tissues because HpaII digests of their exon 4 PCR products showed F₁ pattern but not BALB/c or MSM pattern (data not shown). Subsequent Southern blot hybridization also confirmed these findings (data not shown). The results suggested homozygous deletions at the Ikaros gene region in nine of the 108 lymphomas and that some of the deleted regions may be limited within the coding regions.

We next examined mutations of the Ikaros gene. N-terminal zinc finger domain and activation domain regions were chosen for mutation analysis (Molnar and Georgopoulos, 1994). Figure 4a shows SSCP analysis of PCR products by primers for the activation domain. Bands with aberrant electrophoretic mobility were observed in five of the 108 lymphomas examined. Such band shift was also detected in N-terminal zinc finger domain (data not shown). DNA were recovered from the shifted band areas in the gel and subjected to sequence analysis. Some examples are shown in Figure 4b,c). We identified six and five point mutations in the N-terminal zinc finger domain and the activation domain regions, respectively (Table 1). Among the 11 mutations, five resulted in amino acid-substitutions, two of which changed key amino acid residues of cysteine and histidine in the zinc finger motif. Three were mutations producing stop codons, and the other three comprising one-base insertion or deletion resulted in frameshifts.

The Ikaros and p53 genes reside on the same chromosome 11 in the mouse. This may affect the frequency of tumor development because allelic loss events at a site of chromosome 11 may well involve both loci. Therefore, relation between allelic losses of the Ikaros and p53 loci was investigated. p53 genotyping of lymphomas and the host mice was performed using two sets of primers. One primer set detected the wild-type p53 allele and the other the inserted neo gene (Tsukada et al., 1993; Matsumoto et al., 1998). The typing allowed us to classify 191 lymphomas into four p53 genotypes: (1) (+/+): Both wild-type alleles retained. (2) (+/-): One of the two wild-type alleles lost. (3) (KO/+): The KO allele and the wild-type allele were retained in which the presence of point mutations in the latter was not assessed. (4) (KO/-): The KO allele retained but the wild-type allele lost. The number of lymphomas in these four classes were 72, 3, 48, and 68, respectively.

Table 2 summarizes the allele status of Ikaros and p53 loci in the lymphomas. p53(KO/-) lymphomas showed an allelic loss of the Ikaros gene at a frequency as high as 85% (58/68) and the loss had a preference for the BALB/c allele. This is partly due to the p53 and Ikaros genes being located on the same chromosome. Chromosome 11 derived from the BALB/c parent was preferentially lost in lymphomas because it carried the wild-type alleles of p53 and the Ikaros gene. The frequent allelic loss (69%: 33/48) was also observed in p53(KO/+) lymphomas, much more than that (22%: 16/72) in p53(+/+) lymphomas (² value=24, P value: less than 0.001). This high frequency cannot be accounted for by the synteny, because the loss of BALB/c or MSM Ikaros allele did not accompany the loss of the wild-type p53 allele on the same chromosome. Rather, the result suggests an association between the allelic loss of Ikaros and the presence of p53 KO allele in lymphocytes. Though it was not examined whether or not the wild-type p53 allele in p53(KO/+) lymphomas was inactivated, we infer that some of the lymphomas underwent inactivation of the wild-type p53 allele and that cooperative effects on lymphomagenesis exist between p53 function loss and the Ikaros gene inactivation.

Table 1 shows which of the p53 status of lymphoma the 11 point mutations were derived from. Two mutations were found in 16 p53(+/+) lymphomas, two in 30 p53(KO/+) lymphomas, and seven in 52 p53(KO/-) lymphomas. No particular preference of the mutations to the p53(KO/-) or p53(KO/+) genotypes was noted. This might suggest no direct effect of the p53(KO/-) genotype on the inactivation of the Ikaros gene in lymphomas.

Discussion

In this paper, we show homozygous deletions and mutations of the Ikaros gene in -ray-induced thymic lymphomas. Among 108 lymphomas losing one Ikaros allele as judged by allelic loss analysis, nine lymphomas failed to give PCR-amplification for either the exon 4 primers or the exon 7 primers, indicative of homozygous deletion. Of the nine lymphomas, two retained both end sequences of the BAC insert containing the Ikaros gene, suggesting that some deleted regions are limited within the Ikaros gene region. Ikaros constitutes at least three domains, N-terminal zinc finger domain, activation domain and C-terminal zinc finger domain (Georgopoulos et al., 1992; Molnar and Georgopoulos, 1994; Sun et al., 1996). Six and five point mutations were detected in the N-terminal zinc finger domain and the activation domain, respectively. Of the 11 mutations detected, six led to frameshift mutations or nonsense mutations that resulted in truncated proteins of Ikaros. The homozygous deletions and the mutations of one Ikaros allele together with loss of the other allele lead to loss of function of the Ikaros gene, and therefore the results strongly suggest that Ikaros is involved in the development of -ray induced thymic lymphomas.

The mutations in the zinc finger domain were distributed at various sites but those in the activation domain were clustered in two of the three sites with six contiguous cytosine/guanine residues. Of the latter five mutations, two were one-base insertion, and two were base-substitution associated with one-base deletion. All of these were small mutations, some occurring in the characteristic motif. One base deletion/insertion at homopolymer stretches occur frequently by slippage during DNA replication (Friedberg et al., 1995). Resulting mismatch in a stretch of DNA can be corrected by the mismatch repair system (Kolodner, 1996). Radiation is known to induce DNA strand breaks which are repaired by homologous and illegitimate recombination processes. The latter mechanism of repair frequently produces promiscuous rejoining of two DNA ends and it is well accepted that the major type of radiation-induced mutations is so called large mutations such as deletion and translocation (Thacker, 1992). Although homozygous deletion of the Ikaros gene in radiogenic thymic lymphomas is consistent with the mode of action of radiation, small mutations found in the present study contradict with the expected spectrum of radiation induced mutation. Therefore, the role of radiation in the development of thymic lymphomas may not simply be the direct one in which radiation induces carcinogenic mutation on T lymphocytes, but may involve a variety of biological processes such as stimulation of DNA synthesis during thymic regeneration and delayed mutation in the irradiated target cells (Little et al., 1997).

Ikaros is a transcription factor detected in pluripotent hemopoietic stem cells and its expression is maintained at high levels in mature lymphocytes (Georgopoulos et al., 1992; Molnar and Georgopoulos, 1994; Sun et al., 1996). Recent studies of subnuclear localization in B cells, however, have provided further evidence that Ikaros may not be a simple transcriptional activator (Brown et al., 1997; Klug et al., 1998). Co-localization of Ikaros proteins with transcriptionally silent genes at centromeric heterochromatin suggests a role in transcriptional silencing through the recruitment of the genes to the heterochromatic loci (Brown et al., 1997). The Ikaros protein isoforms that are generated by different splicing all share a common C-terminal domain containing two zinc fingers to which different combinations of N-terminal zinc fingers are appended. The N-terminal zinc fingers are required for sequence-specific DNA binding while the C-terminal zinc fingers mediate homo- and hetero-dimerization among the Ikaros isoforms that greatly increases both the affinity to DNA and the transcriptional activity (Sun et al., 1996; Morgan et al., 1997; Hahm et al., 1998; Kelley et al., 1998). Transcriptional activation is mediated by the activation domain (Sun et al., 1996). The activation and C-terminal domains are encoded for by the translated exon located the 3' most of the gene.

There are two distinct strains of Ikaros-KO mice produced (Winandy et al., 1995; Wang et al., 1996). One carries a mutation targeted by a deletion of the Ikaros DNA-binding domain that displays dominant negative effects on transcription through interaction at the C-terminal domain with the Ikaros isoforms (Ikaros-DN type), and the other harbors a deletion of the last translated exon (Ikaros-C type). Ikaros DN-/- mice lack all fetal and adult lymphoid lineages whereas C-/- homozygotes display selective defects in their fetal and adult lymphoid compartments. Interestingly, heterozygous mice of the two strains show different phenotypes of tumor development. Heterozygotes of the Ikaros-DN mouse strain exhibit defects in the T lineage, which lead to an abnormal accumulation of CD4/CD8 double-positive thymocytes and ultimately result in T cell leukemia and lymphomas. The aberrant expansion of thymocytes and mature T cells is explained by that dominant-negative Ikaros isoforms having an intact C-terminal dimerization domain can sequester the DNA-binding isoforms made by the wild-type allele and dramatically reduce Ikaros activity. In contrast, heterozygotes of the Ikaros-C strain are not obviously abnormal probably because of the null phenotype of the mutation (Wang et al., 1996). On the other hand, Ikaros C-/- mice that lack the Ikaros activity show predominance of oligoclonal and monoclonal thymocyte populations in the thymus at older ages. These results suggest that a profound decrease of Ikaros activity in DN+/- mice or lack of Ikaros activity in C-/- mice leads to T cell hyperproliferation, aberrant expansion of thymic clones, and eventually T cell neoplasia. The Ikaros mutations detected in this study are located at the N-terminal zinc finger domain and the activation domain. Most of them result in null mutation and are therefore similar to the C-/- mutation. Homozygous deletions are also mutation of this type. On the other hand, the two base-substitutions leading to amino acid changes of cysteine and histidine in the zinc finger motif may be mutations of DN-type. We do not notice any difference in tumor properties between lymphomas with the two different types of mutation.

Somatic mutation of the Ikaros gene in the lymphomas is probably one of several genetic alterations involved, since carcinogenesis is a multistep process that includes the inactivation of tumor suppressor genes and activation of proto-oncogenes (Vogelstein and Kinzler, 1993). Inactivation of the p53 gene or a putative tumor suppressor gene(s) on chromosome 12 or on chromosome 16 can be one of the other alterations (Matsumoto et al., 1998; Shinbo et al., 1999). It is noteworthy that the allelic loss of Ikaros was found in p53(KO/+) and p53(KO/-) lymphomas much more than in p53(+/+) lymphomas. The high frequency of Ikaros allele loss in p53(KO/-) lymphomas can be explained by the presence of the Ikaros and p53 genes on the same chromosome which tends to result in loss of both Ikaros and wild-type p53 alleles by a single genetic event such as mitotic recombination, non-disjunction or chromosomal deletion. It was proposed that compound heterozygotes in which mutant alleles for both genes are on the same chromosome develop tumors much more rapidly than mice that carry the two mutations on different chromosomes (Takaku et al., 1998). However, the high frequency of the loss in p53(KO/+) lymphomas cannot be explained by this mechanism. It may reflect cooperative function of the p53 and Ikaros losses in the tumor development.

Ikaros controls T cell proliferative responses and probably functions as a growth suppressor, and hence loss of the gene function may confer a growth advantage to these cells which fosters the development of a tumor (Winandy et al., 1995; Wang et al., 1996). This gene function in tumorigenesis may be similar to that of Rb, a negative regulator of cell growth. On the other hand, p53 can induce either G1 cell-cycle arrest or apoptosis (Halevy et al., 1991; Kuerbitz et al., 1992; Hartwell and Kastan, 1994; Lowe et al., 1993; Sherr, 1996; Giaccia and Kastan, 1998). Loss of cell-cycle control by p53 inactivation can provide negative cell growth and genomic instability (Lengauer et al., 1998). Cooperative tumorigenic effects of germline mutations in Rb and p53 have been discussed (Williams et al., 1994; Ghebranious and Donehower, 1998) and the discussion giving three possibilities can be applied to the cooperativity between Ikaros and p53 losses. (1) Loss of both negative growth regulators is necessary for lymphomagenesis. (2) The absence of p53 might lead to an increased rate of the loss or mutation of Ikaros allele due to impairment of maintaining genomic stability. (3) Ikaros mutation might result in a state of abnormal cell cycle regulation in which cells become prone to p53-dependent apoptosis. Elimination of p53 function would allow these cells to escape the apoptosis and progress through tumorigenesis.

The mouse Ikaros gene has been postulated to participate in proliferation of thymocytes and oncogenic process from studies of Ikaros KO mice as described above. However, no genetic evidence has been presented implicating this gene in the development of lymphomas in Ikaros wild-type mice. Besides, the importance of Ikaros inactivation in lymphomagenesis was not addressed. Here, we have provided evidence for the existence of nine homozygous deletions and eleven point mutations in 191 -ray induced thymic lymphomas, most of which lead to the inactivation of the Ikaros gene. The result strongly suggests a direct role of Ikaros in oncogenesis of mouse thymic lymphomas. This study provides the experimental basis for further analysis of Ikaros mutations in human cancer.

Materials and methods

Mice and lymphomas

MSM is an inbred strain derived from Japanese wild mice, Mus musculus molossinus (Bonhomme and Guenet, 1989). The details of lymphoma induction were described previously (Matsumoto et al., 1998). In brief, the parental p53-deficient mouse was originally produced by introduction of a neo-gene fragment into the p53 gene locus in the ES cells that had been derived from F₁ mouse between C57BL/6(B6) and CBA strains (Tsukada et al., 1993). We have developed a partially congenic MSM mouse carrying a p53-deficient allele by back-crossing this p53-deficient mouse to MSM mice 13 generations. The male mice (N₁₀ to N₁₃ generation) with the genotype of p53(KO/+) were mated with BALB/c female mice and 296 F₁ hybrid mice were obtained. The mice were subjected to -ray irradiation, 2.5 Gy four times at weekly intervals, starting at the age of 4 weeks. Development of thymic lymphoma was diagnosed by the inspection of labored breathing. A total of 205 thymic lymphomas and 29 tumors of other types were obtained.

PCR analysis and SSCP

Isolation of genomic DNA from lymphomas and brain was carried out by standard protocols. Polymerase chain reaction (PCR) and separation of PCR products by gel electrophoresis were performed as described previously (Matsumoto et al., 1998). PCR - SSCP analysis was carried out as described (Orita et al., 1989).

Primer sequences

Microsatellite markers were synthesized according to sequences reported (Dietrich et al., 1996). Primers for the catenin gene are as described previously (Sato et al., 1998). The other primer sequences used are as follows: Ikaros exon 4, 5'-TGAGAAGCCCTTCAAATGCC (FE-4) and 5'-CTGGGAACATGGAACACATG (RI-4) for detection of a polymorphism; 5'-GCTCTCTCTCAGTGCTTACC (FI-3) and RI-4 for homozygous deletion analysis and mutation analysis. Ikaros exon 7, 5'-CTATGACAGTGCCAACTATGAG (FE-7a) and 5'-TGATGGCATTGTTGATGGCC (RE-7a) for homozygous deletion; 5'-AGGGAGACAAGTGCCTGTCA (FE-7b) and 5'-CAGCAGCAAGTTATCCACGG (RE-7b) for SSCP and sequencing. BAC-E1, 5'-CTTCAGCTATGAGAAGGAGC and 5'-TAGCACTCACGCCATAACGC. BAC-E2, 5'-TAGATGCTGTGGCAGGTATG and 5'-CAAGGTAAAGATGCACATATTTG.

Isolation of BACs and pulsed-field gel electrophoresis of DNA

BAC clones were isolated by PCR screening of a library that was purchased from Research Genetics, Inc. The size of clones was determined by pulsed-field gel electrophoresis as described (Hayashi et al., 1993). Each end of BAC inserts was directly sequenced and PCR primers for the BAC ends were synthesized accordingly.

Typing of p53

p53 genotyping was carried out as described previously (Matsumoto et al., 1998; Koide et al., 1999). One primer (F1-53) located in exon 1 of the p53 gene, a second primer (R1-53) in a region 5' to exon 3, and the remaining one (F2-neo) in the neo gene insert. F1-53 and R1-53 amplified a region of the p53 gene to produce a fragment of 500 bp. F2-neo and R1-53 gave an 800 bp fragment comprising a part of the neo and p53 genes. Therefore, the 500 bp band and the 800 bp band represent the normal allele and the mutant allele, respectively. This primer set was also used for allelic loss analysis of lymphomas developing in p53(KO/+).

Acknowledgements

We thank Yuko Saito for technical assistance. This work was supported by grants-in-aid for Research on Human Genome and Gene Therapy from the Ministry of Health and Welfare of Japan.

Bonhomme F and Guenet J-L. (1989). In: Genetic Variants and Strains of the Laboratory Mouse. Lyon MF and Searle AG (eds). Oxford University Press: Cambridge, p.658.

Brown KE, Guest SS, Smale ST, Hahm K, Merkenschlager M and Fisher AG. (1997). Cell 91, 845-854. MEDLINE

Dietrich WF, Radany EH, Smith JS, Bishop JM, Hanahan DH and Lander ES. (1994). Proc. Natl. Acad. Sci. USA 91, 9451-9455. MEDLINE

Dietrich WF, Miller J, Steen R, Merchant MA, Damron-Boles D, Husain Z, Dredge R, Daly MJ, Ingalls KA, O'Connor TJ, Evans CA, DeAngelis MM, Levinson DM, Kruglyak L, Goodman N, Copeland NG, Jenkins NA, Hawkins TL, Stein L, Page DC and Lander ES. (1996). Nature 380, 149-152. MEDLINE

Friedberg EC, Walker GC and Siede W. (1995). In: DNA Repair and Mutagenesis. American Society for Microbiology, Washington.,

Georgopoulos K, Bigby M, Wang J-H, Molnar A, Wu P, Winandy S and Sharpe A. (1994). Cell 79, 143-156. MEDLINE

Georgopoulos K, Moore DD and Derfler B. (1992). Science 258, 808-812. MEDLINE

Ghebranious N and Donehower LA. (1998). Oncogene 17, 3385-3400. MEDLINE

Giaccia AJ and Kastan MB. (1998). Genes Dev. 12, 2973-2983. MEDLINE

Hahm K, Cobb BS, McCarty AS, Brown KE, Klug CA, Lee R, Akashi K, Weissman IL, Fisher AG and Smale ST. (1998). Genes Dev. 12, 782-796. MEDLINE

Halevy O, Rodel J, Peled A and Oren M. (1991). Oncogene 6, 1593-1600. MEDLINE

Hartwell LH and Kastan MB. (1994). Science 266, 1821-1828. MEDLINE

Hayashi T, Ohtsuka H, Kuwabara K, Mafune Y, Miyashita N, Moriwaki K, Takahashi Y and Kominami R. (1993). Genomics 17, 490-492. MEDLINE

Kelley CM, Ikeda T, Koipally J, Avitahi N, Georgopoulos K and Morgan BA. (1998). Curr. Biol. 8, 508-515. MEDLINE

Kemp CJ, Fee F and Balmain A. (1993). Cancer Res. 53, 6022-6027. MEDLINE

Klug CA, Morrison SJ, Masek M, Hahm K, Smale ST and Weissman IL. (1998). Proc. Natl. Acad. Sci. USA 95, 657-662. Article MEDLINE

Knudson AG. (1985). Cancer Res. 45, 1437-1443. MEDLINE

Koide N, Matsumoto Y, Kosugi S, Chou D, Sakai K, Hatakeyama K, Niwa O and Kominami R. (1999). Mol. Carcinog. 24, 57-63. MEDLINE

Kolodner R. (1996). Genes Dev. 10, 1433-1442. MEDLINE

Kuerbitz SJ, Plunkett BS, Walsh WV and Kastan MB. (1992). Proc. Natl. Acad. Sci. USA 89, 7491-7495. MEDLINE

Lengauer C, Kinzler KW and Vogelstein B. (1998). Nature 396, 643-649. Article MEDLINE

Little JB, Nagasawa H, Pfenning T and Vetrovs H. (1997). Radiat. Res. 148, 299-307. MEDLINE

Lowe SW, Schmitt EM, Smith SW, Osborne BA and Jacks T. (1993). Nature 362, 847-852. MEDLINE

Matsumoto Y, Kosugi S, Shinbo T, Chou D, Ohashi M, Wakabayashi Y, Sakai K, Okumoto M, Mori N, Aizawa S, Niwa O and Kominami R. (1998). Oncogene 16, 2747-2754. MEDLINE

Molnar A and Georgopoulos K. (1994). Mol. Cell. Biol. 14, 8292-8303. MEDLINE

Molnar A, Wu P, Largespada DA, Vortkamp A, Scherer S, Copeland NG, Jenkins NA, Bruns G and Georgopoulos K. (1996). J. Immunol. 156, 585-592. MEDLINE

Morgan B, Sun L, Avitahl N, Andrikopoulos K, Ikeda T, Gonzales E, Wu P, Neben S and Georgopoulos K. (1997). EMBO J. 16, 2004-2013. Article MEDLINE

Orita M, Suzuki Y, Sekiya T and Hayashi K. (1989). Genomics 5, 874-879. MEDLINE

Sato S, Tunashima K, Mafune Y, Ohtsuka H, Koide N, Endo Y, Saito Y, Asakura H and Kominami R. (1998). Oncogene 17, 105-113. MEDLINE

Sherr CJ. (1996). Science 274, 1672-1677. Article MEDLINE

Shinbo T, Matsuki A, Matsumoto Y, Kosugi S, Takahashi Y, Niwa O and Kominami R. (1999). Oncogene 18, 4131-4136. MEDLINE

Sun L, Liu A and Georgopoulos K. (1996). EMBO J. 15, 5358-5369. MEDLINE

Takaku K, Oshima M, Miyoshi H, Matsui M, Seldin MF and Taketo MM. (1998). Cell 92, 645-656. MEDLINE

Thacker J. (1992). Adv. Radiat. Biol. 16, 77-117.

Tsukada T, Tomooka Y, Takai S, Ueda Y, Nishikawa S, Yagi T, Tokunaga T, Takeda N, Suda Y, Abe S, Matsuo I, Ikawa Y and Aizawa S. (1993). Oncogene 8, 3313-3322. MEDLINE

Vogelstein B and Kinzler KW. (1993). Trends Genet. 9, 138-141. MEDLINE

Wang J-H, Nichogiannopoulou A, Wu L, Sun L, Sharpe AH, Bigby M and Georgopoulos K. (1996). Immunity 5, 537-549. MEDLINE

Weinberg RA. (1991). Science 254, 1138-1146. MEDLINE

Williams BO, Remington L, Albert DM, Mukai S, Bronson RT and Jacks T. (1994). Nature Genet. 7, 480-484. MEDLINE

Winandy S, Wu P and Georgopoulos K. (1995). Cell 83, 289-299. MEDLINE

Figure 1 Allelic loss analysis of the D11Mit62, Ikaros and D11Mit133 loci. PCR products for D11Mit62 and D11Mit133 and HpaII digests of the Ikaros exon 4 PCR products were subjected to gel electrophoresis. The first three lanes on panels represent control DNA of BALB/c, MSM and F₁ mice. The other lanes display lymphoma DNA. The number of lymphomas is given arbitrarily. D11Mit62 and D11Mit133 both showed polymorphic patterns of bands, and HpaII digests of the 146 bp PCR fragment yielded 86 bp and 60 bp bands in BALB/c and 65 bp and 43 bp bands in MSM (two other 21 bp and 13 bp fragments were missing in this gel). Type of allelic losses in each lymphoma (see legend to Figure 2) is indicated at the bottom

Figure 2 Chromosomal constitution of lymphomas in the vicinity of the Ikaros locus. Three loci used for allelic loss analysis are indicated on the left. Genetic distances between D11Mit62 and Ikaros and between Ikaros and D11Mit133 are approximately 1.8 cM and 0.9 cM, respectively. Triangles marked by C and M indicate BALB/c allele and MSM allele retained, respectively. Two triangles of C and M represent both BALB/c and MSM alleles retained, and one triangle represents one allele lost. Combination of the allele status's allows lymphomas to be divided into five types: type A, lymphomas retaining both alleles of all three loci; type B, lymphomas losing one chromosomal region of BALB/c or MSM that covers all three loci; type C, lymphomas showing allelic losses at the centromeric two loci but retaining both alleles of the telomeric locus; type D, lymphomas retaining both alleles of the centromeric locus but showing allelic losses at the telomeric loci; type E, lymphomas exhibiting allelic loss only at the Ikaros locus. The number of lymphomas of each type is listed at the bottom. Nine of the 191 lymphomas examined are not included because they showed homozygous deletion

Figure 3 Homozygous deletions in the Ikaros gene region. (a) a map showing positions of the Ikaros exons and both ends of an approximately 170 kb BAC insert containing Ikaros. (b) gel electrophoresis of DNA fragments that were produced by multiplex PCR. Products for two BAC ends (E1 and E2), exon 4 and exon 7 are indicated by an arrow in each panel and those for the catenin gene marked by a bracket. Lymphomas 4, 5 and 6 show deletions covering at least exons 4 and 7

Figure 4 Mutations of the Ikaros gene in lymphomas. (a) SSCP analysis of PCR products for the activation domain of Ikaros. Lymphomas 1, 3, 6 and 9 yielded bands with aberrant mobilities suggestive of mutations. (b) Sequencing of PCR fragments in the mobility-shifted bands. (1) normal DNA. (2) lymphoma 3 in the SSCP gel; C7 indicates a cytosine insertion leading to frame-shift mutation in one Ikaros allele of the lymphoma as compared to that of normal cell DNA of F₁ mouse. (3) lymphoma 9 in the SSCP gel; the C to T substitution generates a TAG codon resulting in a premature stop. (c) Sequencing of the N-terminal zinc finger domain. The C to G substitution is shown, which results in the amino acid change from histidine to glutamine in the zinc finger motif

Table 1 Mutations of the Ikaros gene

Table 2 Distribution of Ikaros genotypes in lymphomas of four different p53 genotypes