Screening for Gene Mutations in a 500 Kb Neuroblastoma Tumor Suppressor Candidate Region in Chromosome 1P; Mutation and Stage-Speciﬁc Expression in UBE4B/UFD2

Screening for gene mutations in a 500 kb neuroblastoma tumor suppressor candidate region in chromosome 1p; mutation and stage-speciﬁc expression in UBE4B/UFD2

Cecilia Krona1, Katarina Ejeska¨ r1,2, Frida Abel1, Per Kogner3, Jill Bjelke1, Elin Bjo¨ rk1, Rose-Marie Sjo¨ berg1 and Tommy Martinsson1

1Department of Clinical Genetics, Institute for the Health of Women and Children, Go¨teborg University, Sahlgrenska University Hospital-East, SE-41685 Go¨teborg, Sweden; 2Cell and Gene Therapy Group, Murdoch Children’s Research Institute, Royal Children’s Hospital, Parkville, Vic. 3052, Australia; 3Childhood Cancer Research Unit, Karolinska Institute, Karolinska Hospital, SE-17176 Stockholm, Sweden

Deletion of a part of the short arm of chromosome 1 is one adrenal medulla (as reviewed by Brodeur, 1990). It is the of the most common chromosomal rearrangements third most common pediatric cancer and the one most observed in neuroblastoma (NBL) tumors and it is often seen in newborns (reviewed in Maris and Matthay, associated with a poor prognosis. No NBL tumor 1999). NBL is characterized by vast heterogeneity, suppressor gene has yet been identified in the region. ranging from spontaneous regression to aggressive Our shortest region of overlap of deletions, ranging from malignant behavior often with a fatal outcome. A marker D1S80 to D1S244, was shown to partly overlap a particularly aggressive subtype of NBLs is genetically 500 kb region that was homozygously deleted in a NBL characterized by amplification of the MYCN oncogene cell line. We have screened seven genes known to reside in and of chromosome 1p deletions (for a review, see Maris or very close to this overlap consensus region, UBE4B/ and Matthay, 1999). Gain of parts of the long arm of UFD2, KIF1B, DFFA, PGD, CORT, PEX14,andICAT, 17q (17q gain) has also been shown to be a characteristic for coding mutations in NBL tumor DNA. A few feature of the aggressive subtype (Lastowska et al., 1997; deviations from the reference sequences were identified; Abel et al., 1999; Bown et al., 1999). The frequent 1p- most interestingly being a splice site mutation that was deletion in clinically advanced tumors implies that the detected in UBE4B/UFD2 in a stage 3 NBL with a fatal 1p-region contains at least one gene (denoted TSG; outcome. This mutation was neither present in the patients tumor suppressor gene; NBS, OMIM 256700) that is constitutional DNA nor in any of 192 control chromo- prone to inactivation in NBL tumors. somes analysed. Also, the expression of UBE4B/UFD2 Our group and others have tried to narrow down the was markedly diminished in the high-stage/poor-outcome consensus region of the NBSlocus (Caron et al., 1995, tumors as compared to the low-stage/favorable-outcome 2001; Maris et al., 1995, 2001; Martinsson et al., 1995, tumors. Overall, the number of amino-acid changes in the 1997; White et al., 1995, 2001; Ichimiya et al., 1999; genes of the region was low, which shows that mutations in Bauer et al., 2001; Ejeska¨ r et al., 2001; Spieker et al., these genes are rare events in NBL development. Given 2001). Our previous efforts to delimit the 1p-region most the data presented here, UBE4B/UFD2 stands out as the commonly deleted in NBL tumors, primarily of strongest candidate NBL tumor suppressor gene in the Scandinavian origin have identified a shortest region region at this stage. of overlap (SRO) of deletions in our material of Oncogene (2003) 22, 2343–2351. doi:10.1038/sj.onc.1206324 approximately 25 cM between distal marker D1S80 and proximal marker D1S244 on the short arm of Keywords: 1p36.2-3; neuroblastoma; mutation analysis; chromosome 1 (Martinsson et al., 1995, 1997). Subse- embryonal tumors quent 1p-LOH studies of both NBLs and germ cell ONCOGENOMICS tumors identified a 5 cM combined SRO of deletion for these tumor types defined distally and proximally by markers D1S508 and D1S244, respectively (Ejeska¨ r Neuroblastoma (NBL) is a tumor of the sympathetic et al., 2001). nervous system arising from the neural crest. It is Further support for the localization of a TSG to this therefore composed of undifferentiated neuroectoder- region was obtained when a homozygously deleted mal cells and the most common site of localization is the region at 1p36.2-3 was found in an NBL cell line (Ohira et al., 2000). This deleted region maps in the proximal part of the NBL consensus deletion region defined by us, *Correspondence: T Martinsson; close to the D1S244 marker, and it is delimitated by the E-mail: [email protected] Received 18 July 2002; revised 10 December 2002; accepted 11 December genes UBE4B/UFD2 (ubiquitination factor E4B – 2002 homologous to yeast gene) and PEX14 (peroxisomal Neuroblastoma tumor suppressor candidate region CKronaet al 2344 biogenesis factor 14; OMIM 601791), with KIF1B alpha; OMIM: 601882), PGD (phosphogluconate dehy- (kinesin superfamily protein member 1B; OMIM: drogenase; OMIM: 172200) and CORT (cortistatin, 605995), DFFA (DNA fragmentation factor, 45 kDa, OMIM 602780) in between (Figure 1a). Also, located

a 9.4 9.5 9.6 9.7 9.8 9.9 Mb from 1pter

Ohira et al., 2000

MartinssonMartinsson et et al., al., 1995 1995 D1S244

UBE4B KIF1B PGD DFFA PEX14 9.2 CORT

ICAT

b ; np) >T C ; > C; C C C np/ > > 80C

UBE4B 1G splicesite L571F( 2952T> 2985T 3909+ Q1075H(p/b) 3594A>G 3225G> 1713G 1439+

5' 1 2 3 4 5 6 7 8 9 10 11... 17 181920 21 22 23 24 25 26 2728 3'

38.4 23.3 5.6 2.6 10.9 2.0 4.7 1.0 1.5 1.9 1.9 2.1 2.7 2.9 0.6 UTR 1.9 0.5 2.4 2.5 0.1 2.5 7.7 2.1 6.8 6.9 7.3 0.3 UTR

PGD >A D246N 120C>T acidic/up 736G>A; 1437G

5' 1 2 3 4 5 6 7 8 9 10 11 12 13 3'

0.5 0.7 2.5 1.0 3.8 3.3 1.5 3.7 0.3 1.3 0.5 0.1 UTR UTR T T C > > 1014C PEX14 156T > 768G >A 1032G

5' 1 2 3 4 5 6 7 8 9 3'

20.3 41.0 63.0 19.1 4.7 1.3 3.0 2.2 UTR UTR

ICAT 5' 1 2 3 3'

5.7 5.9 0.8 20.5 UTR UTR UTR UTR

Oncogene Neuroblastoma tumor suppressor candidate region C Krona et al 2345 just distal of this region is the ICAT (beta-catenin chaperone for DFF40 (Sakahira et al., 1999, 2000). In interacting protein) gene. Several of these genes are two tumors one rare allele variant, causing a nonpolar likely to be involved in cellular proliferation and/or to an uncharged polar amino-acid exchange (I69T), differentiation and may constitute excellent candidates could be detected in the coding region of DFFA. This for being a tentative NBSgene. We have previously region, where the exchange occurs, is predicted to be presented data on the CORT and DFFA genes involved in the binding between DFF45 and DFF40 performed on a set of NBL tumors (Ejeska¨ r et al., (Abel et al., 2002) This amino-acid substitution might 2000; Abel et al., 2002). disrupt the ability of DFF45 to help DFF40 to acquire In this paper, we have analysed the remaining five its final, appropriately folded form. Cells with this genes in the region in a set of NBL tumors. The mutation can consequently have a reduced ability to complete coding region of UBE4B/UFD2, PGD, induce apoptotic mechanisms such as DNA fragmenta- PEX14, and ICAT have now also been determined in tion and chromatin condensation. a set of NBL tumors using direct DNA-sequencing, Selective protein degradation through the ubiquitin/ together with most of the KIF1B gene. Thus, we have in proteasome system plays an important role in eukaryote detail studied the genes in the combined region as cellular regulation. It is also essential for protecting cells suggested by Ohira et al. (2000) and by us (Martinsson against environmental stress because of its role in the et al., 1995) in a tumor set separate from that studied by elimination of aberrant proteins generated under normal Ohira et al. Also, the expression of the genes in NBL and, in particular, stress conditions (Hochstrasser, tumors was analysed. Several new polymorphisms in the 1996). Efficient multi-ubiquination needed for protea- genes are presented, both in coding and in noncoding somal targeting of a protein requires the conjugation regions. One true mutation in the UBE4B/UFD2 gene factor E4 of which family UBE4B/UFD2 is a member was detected in a stage 3 NBL with a fatal outcome. (Koegl et al., 1999). In yeast, the E4 homologue, UFD2, This fact and the very low expression of UBE4B/UFD2 is linked to cell survival under stress conditions. It is in high-stage tumors are indicative of UBE4B/UFD2 as likely to mediate the degradation of stress-induced a player in NBL development and/or progression. aberrant proteins. This makes UBE4B/UFD2 a con- DFFA is a gene coding for the protein DFF45 and the ceivable candidate TSG based on its functional char- short splice variant DFF35. Mutation analysis data for acteristics as well as the position of the gene in a NBL this gene have been presented earlier by our group (Abel hot spot region. The gene structure of UBE4B/UFD2 as et al., 2002). DFF45 must be cleaved by caspase-3 and derived from the alignment of the reference sequence dissociate from the DFF40/DFF45 complex for the 40- NM_006048 with the UCSC (URL: http//genome.ucs- kDa subunit of DNA fragmentation factor to induce c.edu) chromosome 1 draft sequence is shown in DNA fragmentation and chromatin condensation dur- Figure 1b. The gene consists of a coding sequence of ing apoptosis (Liu et al., 1997, 1998). The full-length 3909 bases distributed over 28 exons. The short exon 1 variant of DFF45 also has an important function as a sequence was not aligned to the UCSC draft sequence

Figure 1 (a) Map of the genes in the analysed region. The location of the genes are based on data from the UCSC genome browser; April 2002 draft sequence (URL: http//genome.ucsc.edu. The homozygously deleted region on chromosome 1p in a NBL cell line, found by Ohira et al. (2000), is indicated by a dark shaded box, while the NBL SRO as defined by our group (Martinsson et al., 1995) is shown as a light gray box. The region includes the genes UBE4B/UFD2,KIF1B,DFFA,PGD,CORT and PEX14. ICAT is located just distal of the region. Arrows indicate the transcriptional direction of the genes. The scale in the top is the approximate distance from the 1p-terminal, in megabase pairs. The study was performed on tumor samples obtained from patients with the diagnosis NBL mainly from the Scandinavian countries. The tumors were staged according to the International NBL Staging System criteria (INSS; Brodeur et al., 1993, 1988). Clinical data on patients have been published earlier (Martinsson et al., 1995; Abel et al., 2002). Genomic DNA was extracted from frozen (À701C) tumor tissue, and in some cases also from corresponding peripheral blood, according to standard procedures. (b) Genomic structures of the genes UBE4B/UFD2,PGD,PEX14, and ICAT with schematic representations of exon– intron structures together with the detected deviations from the reference sequences. Shaded boxes represent the coding regions and dark boxes represent the 50- and 30-untranslated regions of the genes. The intron sizes are indicated in the figure in kilobase pairs. Information on sequences for PCR primers used, annealing temperatures, and data on the number of analysed tumor and control samples for each exon are available on request. Initially, 23 tumors of all different stages were chosen for sequencing of UBE4B/UFD2, PEX14,PGD, and ICAT. Stages 1, 2a, and 4S were represented by one tumor each, while advanced stages 3 and 4 were represented by 10 and 9 tumor cases, respectively. After this initial screening some exons that showed deviations relative to reference sequence was further analysed in an additional set of tumors. For several exons, approximately 40–100 tumor samples were analysed in total (Tables 1 and 2). Control samples, derived from normal blood, were included in the sequencing of all exons. For some exons, where deviations were detected, an additional number of controls were sequenced (see Tables 1 and 2); in confirmation cases approximately 100 control samples were analysed. An initial screening was performed on most of the exonic sequences from the KIF1B gene. Since this revealed no variations from the reference sequence, further analysis of this gene was not undertaken (data not shown). The changes found in the coding sequences of the genes by DNA sequencing are denoted by lines connecting the sequence deviations to the corresponding sites in the figure derived by in silico mapping using the respective reference sequences. Changes in coding sequences are written as the example (1713G>C; L571F) with the A of the initiator Met codon denoted nucleotide +1. The amino acids are represented by their single letter code with the amino acid in the reference sequence to the left and the alternative amino acid, deduced from the translation of the sequenced fragment, to the right with the position of the substitution in the protein sequence in between. In the cases where the change in nucleotide sequence is not silent at protein level, the type of amino-acid change is marked; np, nonpolar; up, uncharged polar; b, basic. 1439+1G>C represents a mutation in the donor splice site of exon 9 in UBE4B/UFD2. The polymorphism in the 3’UTR region, 80 bases after exon 28 of UBE4B/UFD2, has a previous record in dbSNP (rs1046277). The polymorphism in exon 3 of P6D has two previous records in dbSNP (rs49887/2070193) (PEX14)

Oncogene Neuroblastoma tumor suppressor candidate region CKronaet al 2346 when NM_006048 was used for the BLAT search. Sequence analysis identified a few polymorphic Instead, the cDNA AB028839 was used to define the changes in UBE4B/UFD2 (Tables 1 and 2). One of boundaries of these 24 nucleotides after the initiation them, which was located in the 30UTR of the gene codon. (c.3909+80T>C; where c.1 represents the first base of

Table 1 Observed changes in coding sequences Gene Exon Change in Predicted Type Affected cases Frequency Frequency Case numbers number coding change in among among sequence protein patients controls UBE4Ba 9 c.1439+1G>Cb Aberrant Unknown Heterozygous G/C 1/89 0/96 St131 splicing Homozygous C/C 0/89 0/96 13 c.1713G>C L571F np/np Heterozygous G/C 2/94 2/90 FH177, FH185 Homozygous C/C 0/94 0/90 22 c.2952T>C S984S No aa change Heterozygous T/C 11/60 1/1 FH177, St119, St172, FH146, St129B, St234, St190, St214B, FH133, FH185, St173B Homozygous C/C 4/60 0/1 St114, FH136, St121, St196 (HemZ?) 22 c.2985T>C H995H No aa change Heterozygous T/C 2/60 0/1 St114, St119 Homozygous C/C 1/60 0/1 St121 (HemZ?) 24 c.3225G>C Q1075H up/basic Heterozygous G/C 3/86 2/93 St234, St172, FH146 Homozygous C/C 0/86 0/93 26 c.3594A>G T1198T No aa change Heterozygous A/G 2/56 0/1 St102, St114 Homozygous G/G 1/56 0/1 St129B

PGDc 3 c.120C>Td D40D No aa change Heterozygous C/T 9/21 1/1 St100, St149, FH125, St116, St119, St153, St216, St223, St235 Homozygous T/T 9/21 0/1 FH146, FH177, FH85, FH157, St99, St114, St124, St102, St172 5 c.403C>Ge P135A np/np Heterozygous C/G 0/21 0/1 Homozygous G/G 21/21 1/1 FH146, FH177, FH85, FH157, St99, St100, St114, St124, St149, St158, St208, FH125, FH169, St102, St115, St116, St119, St153, St172, St216, St223 7 c.546T>Ce P182P No aa change Heterozygous T/C 0/23 0/1 Homozygous C/C 23/23 1/1 FH146, FH177, FH85, FH157, St99, St100, St114, St124, St131, St149, St158, St208, FH125, FH169, St102, St115, St116, St119, St153, St172, St216, St223, St235 8 c.736G>A D246N Acidic/up Heterozygous G/A 8/48 13/28 FH177, FH156, FH51, St214, St107, FH137, St104, FH115 Homozygous A/A 2/48 1/28 FH128, St113 13 c.1437G>A S479S No aa change Heterozygous G/A 1/20 0/1 FH177 Homozygous A/A 0/20 0/1

PEX14f 3 c.156T>Cg F52F No aa change Heterozygous T/C 5/22 1/1 St114, St149, St116, St153, St235 Homozygous C/C 16/22 0/1 FH146, FH177, FH85, FH157, St99, St100, St124, St131, St208, FH125, FH169, St102, St115, St172, St216, St223, 9 c.768G>A V256V No aa change Heterozygous G/A 1/23 0/1 St100 Homozygous A/A 0/23 0/1 c.1014C>Th D338D No aa change Heterozygous C/T 2/23 0/1 St114, FH125 Homozygous T/T 0/23 0/1 c.1032G>T G344G No aa change Heterozygous G/T 3/22 0/1 FH169, St102, St153 Homozygous G/G 0/22 0/1

Column 3, change in coding sequence with the A of the initiator Met codon denoted nucleotide +1; np, column 5, nonpolar; aa, amino acid; up, uncharged polar a Position in NM 006048 with the A of the initiator Met codon (in position 86 of the cDNA) denoted nucleotide +1 b Change from G to C in the donor splice site c Position in NM 002631 with the A of the initiator Met codon (in position 7 of the cDNA) denoted nucleotide +1 d Published dbSNP:rs1049887/2070193 e Probably sequencing errors in the reference sequence NM 002631 f Position in NM 004565 with the A of the initiator Met codon (in position 6 of the cDNA) denoted nucleotide +1 g Published dbSNP:rs12375 h Published dbSNP:rs2128414

Oncogene Neuroblastoma tumor suppressor candidate region C Krona et al 2347 Table 2 Observed variations in noncoding sequences Gene In proximity Variation in Affected cases Frequency Frequency Case numbers to exon noncoding among among number sequence patients controls UBE4B 5 c.436À55C>T Heterozygous C/T 1/66 0/1 FH133 Homozygous T/T 1/66 0/1 St121 14 c.1911+13A>T Heterozygous A/T 32/78 0/1 FH177, FH127, St100, St158, FH162, FH125, St172, FH112, FH114, FH32, FH189, FH84, FH146, FH185, FH51, FH104, FH138, FH85, FH153, St114, St189, FH115, St104, FH126, FH151, St118, FH174, St173, FH176, St121, St130, St214, Homozygous T/T 9/78 0/1 St99, St108, FH136, St131, St142, St116, St190, St113, St234 19 c.2591+47C>A Heterozygous C/A 0/58 0/1 Homozygous A/A 1/58 0/1 St121 22 c.2927À37A>G Heterozygous A/G 11/60 1/1 FH177, St119, St172, FH146, St129B, St234, St190, St214B, FH133, FH185, St173B Homozygous G/G 4/60 0/1 St114, FH136, St121, St196 (HemZ?) 28 c.3909+80T>Ca Heterozygous T/C 4/44 1/1 St190, FH185, St234, St214B Homozygous C/C 39/44 0/1 FH161, FH156, FH127, FH133, St156, FH187, FH128, FH135, FH162, St142, FH51, FH114, FH41, FH32, FH189, FH155, FH95, FH106, FH184, St129, St174, St196, FH84, FH118, St170, St168, St130, St127, St126, St121, FH176, St117, St113, St107, FH137, St109, St104, FH115, St189

PGD 2 c.84+18G>A Heterozygous G/A 8/22 4/8 FH177, St100, FH125, FH169, St116, St153, St223, St235 Homozygous A/A 5/22 0/8 St131, St158, St208, St149, St119 13 c.1333À9G>A Heterozygous G/A 9/20 1/12 FH177, FH85 St100, FH125, FH169, St116, St153, St216, St235 Homozygous A/A 3/20 0/12 St131, St158, St208

PEX14 2 c.37À74C>T Heterozygous C/T 6/23 1/1 St114, St149, St102, St116, St153, St235 Homozygous T/T 15/23 0/1 FH146, FH177, FH85, FH157, St99, St100, St131, St158, St208, FH125, FH169, St115, St172, St216, St223 5 c.384+14A>T Heterozygous A/T 4/23 1/1 FH125, FH169, St153, St235 Homozygous T/T 2/23 0/1 FH146, St208

Column 2, exon from which the position in the intronic sequence is counted; Column 3, variation in noncoding sequence n nucleotides before (À)or after (+) the designated position in the cDNA. Positive numbers starting from the G of the donor site invariant GT, negative numbers starting from the G of the acceptor site invariant AG a Published: dbSNP:rs1046277

the initiation codon and bases after a ‘À’ or a ‘+’ sign A somatic heterozygous mutation was observed in the denote nucleotide distance upstream or downstream, splice donor site of exon 9 (c.1439+1G>C) in 1/89 respectively, from that base in the genomic sequence; tumors (Case St131, a stage 3 tumor with a fatal nomenclature described by Antonarakis, 1998), has outcome). This change from the invariant bases GT to previously been reported in the NCBI dbSNP CT was neither found in the normal DNA of this patient (rs1046277). One conservative amino-acid substitution nor in any of the 192 control chromosomes analysed was observed in exon 13 (c.1713G>C; L571F) in (Figure 2), and thus constitutes a true mutation. This heterozygous form in 2/94 tumors. This variation was change of the evolutionary conserved splice site in one also observed in 2/90 control individuals. Two nucleo- patient is likely to elicit erroneous intron inclusion. A tide transitions were observed in exon 22. Both were premature termination codon is then introduced only a silent at protein level and the ﬁrst (c.2952T>C; S984S) few bases after exon 9, and the function of the gene will was observed also in the control in heterozygous form. most likely be abolished since the C-terminal domain is Another silent transition in exon 26 (c.3594A>G; highly conserved between proteins of the UFD2 family T1198T) was detected in 3/56 tumors, one of them in in different species (Figure 2b, c). It is postulated that homozygous form and the other two in heterozygous binding of UFD2 proteins to ubiquitin–protein con- form. jugates is mediated by this conserved domain (Koegl

Oncogene Neuroblastoma tumor suppressor candidate region CKronaet al 2348 a c exon 9 intron 9 AC AC AG C C C AG G TA T G A A 3,000 2,000 Case St131; Tumor G/C 1,650 1,650 1,000 850 1,000 650 850 G 500 A C A C A G CCCA G C T A T G AA 400 300 650 200 Case St131; Normal G/G 100 500 400

A C A C A G CCCA GGT A T G A 300

200 Normal control G/G

A C A C A G CCCA GGT A T G 100

b 2453 bp Fp Rp

Intron 9 - 2349 bp Exon 9 - 101 bp Exon 10 - 115 bp

58 bp 46 bp

Normally spliced cDNA product 58+46 bp = 104 bp

Non-spliced abberant cDNA product/genomic DNA product 2453 bp

Figure 2 Splice site mutation detected in UBE4B/UFD2.(a) The (c.1439+1G>C) mutation that occurs in the donor splice site of exon 9 in tumor St131 was not found in the constitutive DNA of the patient or in any of 192 control alleles sequenced. The dotted line denotes the border between exon 9 and intron 9. The donor splice site of exon 9 is underlined. Upper panel: amplified tumor-DNA from case St131, heterozygous G/C. Middle panel: amplified constitutional DNA from case St131 (extracted from blood), homozygous G/G. Lower panel: amplified constitutional DNA from control Q392 (from blood), homozygous G/G. For mutation, analysis was performed as follows: primers for the coding sequences of the genes were designed on the basis of the in silico mapping and the UCSC Genome Browser, December 2001 chromosome 1 draft sequence. They were purchased from Life Technologies (www.invitrogen.com; Life Technologies, Inc. Gaithersburg, MD, USA). For DNA-sequencing, the PCR products were purified with ExoSAP-IT (USB, Corp., Cleveland, OH, USA). The sequencing reactions were made using the ABI PRISM Big Dye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems, Foster City, CA, USA) according to ABI protocols. Sequencing was performed using an ABI PRISM 3100 DNA Sequencer (Applied Biosystems). Sequences derived from the tumor samples were compared to the corresponding sequence from DNA of a normal blood donor or a reference sequence derived from NCBI (URL: http:// www.ncbi.nlm.nih.gov. In the cases where a deviation from the reference was observed, it was confirmed through resequencing from the opposite direction. (b) Schematic representation of amplification products from cDNA samples amplified with primers located in exons 9 and 10, respectively. One short fragment is amplified if the transcript is normally spliced, while a much longer fragment is amplified if splicing of the exon 9 donor site to the exon 10 acceptor site does not occur and intron retention is the result. (c) Left panel: amplification of cDNA extracted from blood donors (lanes 1 and 2) and from tumor-DNA from case St131 (lane 3) with the primer pair located as shown in Figure 3b. The fourth lane is a negative control sample, and the fragment visible on the gel in that sample as well as in the blood sample in lane 1 is primer dimers. The other amplification products are consistent with the size of the spliced and the nonspliced fragments indicated in Figure 3b. Right panel: amplification of cDNA extracted from tumor-DNA from case St131 with primers located in the b-actin gene. Amplification of cDNA results in a product size of 284 bp. If genomic DNA had been present in the sample, amplification would have produced a fragment of 397 bp

et al., 1999). However, it is also possible that this point Q1075H), causing an uncharged polar amino acid to mutation, changing GT to CT, will activate a cryptic be replaced by a basic amino acid, was observed in splice site or cause the exclusion of the entire exon 9 heterozygous form in 3/86 tumors. It was also detected together with introns 8 and 9 if the spliceosome fails to in heterozygous form in 2/93 control individuals. Even recognize this sequence as an exon. In exon 24 a though this variation could represent a polymorphism, nonconservative germ line deviation (c.3225G>C; we cannot rule out the possibility that the control

Oncogene Neuroblastoma tumor suppressor candidate region C Krona et al 2349 individuals with this variant allele have a genetic 1997; Miller et al., 1999; Peifer and Polakis, 2000). By its predisposition for tumor development. Linkage disequi- function as a negative regulator of Wnt signaling, librium was present between the three variations in exon inactivation of ICAT could cause inappropriate activa- 22 of UBE4B/UFD2. Variations (c.2927À37A>G) and tion of the target genes in the signaling pathway. ICAT (c.2952T>C; S984S) occurred in the same alleles. The could therefore act as a TSG in the development of the variation c.2985T>C (H995H) in exon 22 was also embryonic cells that eventually will become NBL tumor inherited together with the two former variations in cells. Alignment of ICAT mRNA NM_020248 to the three cases. UCSC chromosome 1 draft sequence displayed a gene The gene structure of PGD as a result of a comparison with three exons in the coding region, which in total is between the reference sequence NM_002631 and the 246 bp long. Figure 1b shows the gene structure of cDNA BC000368 is shown in Figure 1b. As in the case ICAT. Mutation analysis performed for ICAT exons 1– with UBE4B/UFD2, a cDNA with a longer 50UTR than 3 disclosed no deviations from the control DNA or from the reference sequence had to be used to make the the NCBI reference sequence in our set of NBL tumors. alignment of the first bases of the coding sequence of We have analysed CORT in detail earlier (Ejeska¨ r PGD to the UCSC draft sequence possible. The coding et al., 2000) and no mutations were detected in this gene. sequence of the gene consists of 1458 bases distributed Also, KIF1B sequences analysed revealed no significant over 13 exons. One striking finding from the in silico changes. Thus, we could find no evidence that any of mapping of PGD was that the donor splice site of exon 6 these genes are involved in NBL tumorigenesis. was GC instead of the consensus GT in all sequenced RT–PCR expression analysis of UBE4B/UFD2 was samples including the control samples. This is also the performed and compared with RT–PCR of the genes case in the reference sequence NM_002631. The DFFA and GAPDH from the same tumor samples disposition of a noncanonical splice site, that is a GC– (Figure 3a). The expression study of UBE4B/UFD2 in AG pair instead of the canonical dinucleotides GT and NBL tumors followed the same expression pattern as for AG as donor and acceptor sites, respectively, is well DFFA with a markedly lower expression in the high characterized and has been estimated to occur in 0.5– stage tumors. Our study regarding the expression status 1.1% of the total splice site pairs in the genome (Burset of exon 7 in UBE4B/UFD2 clearly indicated that the et al., 2000; Thanaraj and Clark, 2001; Clark and cDNA AB028839 is correct with respect to the sequence Thanaraj, 2002). Sequence analysis demonstrated seven corresponding to exon 7 in NM_006048. Thus, exon 7 is deviations from the reference sequence NM_002631 not included in the transcript in the cases analysed (Tables 1 and 2). All were either polymorphic changes or (Figure 3b). Our observation of reduced expression errors in the reference sequence. No nonpolymorphic patterns of DFFA and UBE4B/UFD2 in high-stage changes were observed in PGD, which indicates that this tumors thus confirms the study by Ohira et al. (2000), in gene is not involved in the pathogenesis of NBL. This is which they observed a significantly higher expression in accordance with the fact that PGD does not have any level of these genes in favorable stage 1 tumors than in characteristics of a classical TSG. unfavorable tumors with MYCN amplification. In silico mapping of PEX14 by alignment of the What are the evidences for a TSG in the region? Even reference sequence NM_004565 to the UCSC chromo- though all the known genes in the region found to be some 1 draft sequence displayed a gene consisting of homozygously deleted in a NBL cell line have been nine different exons (Figure 1b). The gene extends over a screened for mutations, there is still a possibility that large genomic area (approximately 150 000 bp) and thus there are genes in the region that are not yet annotated. has large introns. The mutation analyses of PEX14 It should also be noted that we have primarily analysed revealed four variations in the coding sequences, as well coding regions of the genes in this study. The promoter as two alterations in intronic sequences close before and regions of the genes have neither been sequenced nor after exons. All coding alterations occurred in the third investigated for epigenetic inactivation by promoter base of a coding triplet and they did not cause any methylation. Also, intronic sequences of the region, the amino-acid exchange. An overview of the results from first exon of PEX14 and some of the exons of KIF1B mutation analysis of PEX14 is shown in Tables 1 and 2, were not investigated. However, it does not seem likely and in Figure 1b. Linkage disequilibrium was observed that all of the inactivating mutations of a gene should be between polymorphisms in PEX14 where c.156T>C; located in regulating sequences. If one of the genes in the F52F in exon 3 was shown to link to the variation region was involved in the development of NBL tumors, c.37À74C>T, located 74 bases upstream of exon 2 and we would have expected to observe a higher frequency approximately 50 kb from exon 3. of inactivating mutations in the coding sequence. The ICAT gene, which is located just distal to the At least two alternative NB tumor suppressor loci homozygously deleted region but within the NBL SRO, have been suggested in the 1p35–36 region. LOH studies was screened for mutations based on the candidate of tumors from patients in the Netherlands and Belgium properties of the gene. ICAT may, according to Tago have demonstrated distinct types of SRO in tumors with et al. (2000), interfere with the interaction between beta- or without MYCN amplification (Caron et al., 1995, catenin and TCF-4, and thereby act as a negative 2001). MYCN-amplified NBLs were associated with regulator in the Wnt signaling pathway. This signaling large 1p deletions extending from 1p35–36.1 to the has an important role in both embryonic development telomere. In the MYCN single copy tumors, however, and tumorigenesis (for reviews, see Cadigan and Nusse, the SRO was restricted to 1p36.2–3 and the lost alleles

Oncogene Neuroblastoma tumor suppressor candidate region CKronaet al 2350 a 2001) that a deletion of one allele might be supplemen- Stage: ted by epigenetic mechanisms on 1p36.2–3 that inacti- GN 1 4S 3 4 + vate the remaining TSG allele. Also, other investigators UBE4B have suggested alternative TSG locations (Ichimiya et al., 1999; Bauer et al., 2001; Maris et al., 2001; White DFFA et al., 2001). Therefore, it is a possibility that the homozygously deleted region in a NBL cell line, is a GAPDH secondary effect of tumorigenesis and that an NBL TSG is located more distal on 1p. In conclusion, no solid evidence supporting the b I:1 I:2 I:3 I:4 I:H O II:1 II:2 II:3 II:4 II:H O hypothesis of any of the seven investigated genes in this 2 2 500 kb region on 1p36.2 functioning as an NBL TSG was found. Still, there are a number of intriguing 1000 sequence changes, which together with the reduced expression pattern of these genes (DFFA and UBE4B/ 400 300 UFD2) raise the question of whether any or both of 200 these genes are involved in a pattern of gene disruptions

100 and transcriptional regulation on 1p36.2–3 that colla- borates in the task of tumor development or further progression along the path of malignant changes from normal cellular behavior through unknown mechan- Figure 3 (a) RT–PCR expression analysis of UBE4B/UFD2 and isms. DFFA from all different stages of NBL. GN, ganglioneuroma; 1, We would, however, like to stress that several lines of 4S, 3, and 4, stages of NBL; +, positive control. Upper panel: tumor extracted cDNA amplified by RT–PCR with UBE4B/UFD2 evidence support a role for UBE4B/UFD2 in the NBL cDNA primers. Middle panel: tumor extracted cDNA amplified by development. (i) It is deleted in all NBL cases with a RT–PCR with DFFA primers. Lower panel: GAPDH control deletion as tested by us, that is, it is located in the SRO cDNA amplified by RT–PCR with GAPDH primers (DFFA and region. (ii) It is disrupted by homozygous loss in the GAPDH primers are available on request.). From left to right: GN, case St151; stage 1, FH118, FH161; stage 4S, FH125, FH162; stage NBL cell line case presented by Ohira et al. (2000). (iii) 3, FH157, FH187, St131, St149; stage 4, FH32, FH114, FH126, We have detected a mutation in one primary stage 3 FH168, FH169, FH111, FH106, FH155, FH174, St102, St129, NBL tumor with a fatal outcome. (iv) The function of St130; positive control, cDNA from fibroblasts. Expression UBE4B/UFD2 is well in line with that of a tumor analysis of UBE4B/UFD2 was performed with a selection of the suppressor gene (see discussion above). (v) UBE4B/ tumor panel used in the DFFA study (Abel et al., 2002). This was done in order to make a comparison between the expression UFD2 has a marked stage-specific expression pattern, in patterns of the two genes in the tumor samples. The expression of general with a very low or even absent expression in UBE4B/UFD2 was tested by reverse transcriptase–PCR (RT–PCR) high-stage/poor outcome tumors relative to the strong with a forward primer located in exon 3: 50-CTCTCAG- expression of the low-stage/favorable outcome tumors. CAGCTCGCCCTCTA-30 and a reverse primer in exon 5: 50- TTCCTGAATCCACATCCACCTG-30. This generates a cDNA (vi) An additional interesting fact is that a strongly fragment of 126 bp and, theoretically, a genomic DNA fragment of homologous gene to UBE4B/UFD2, that is, UBE4A,is approximately 6000 bp. (b) RT–PCR expression analysis of exon 7 one of a few genes mapped to the 2–5 cM consensus 11q- in UBE4B/UFD2. The expression of the sequence corresponding to deletion region shown in a group of NBLs (Guo et al., exon 7 in NM_006048 was investigated by using two primer pairs, I 1999, 2000; John M Maris, personal communication). and II, located in exons 6 and 8 and cDNA from tumors 1–4. Tumors 1 and 2 are stage 1 NBLs, tumor 3 is a stage 3 NBL and This group that consists of NBLs without MYCN tumor 4 is a stage 4 NBL. The expected sizes of the PCR products amplification and with a poor prognosis could be a are 206 and 311 bp for primer pairs I and II, respectively, assuming ‘complementary’ group to those discussed in this paper; that exon 7 is not included in the transcript. If exon 7 had been 1p-deletion generally goes with MYCN amplification in expressed, PCR products of 593 and 698 bp would have been amplified with the same primer pairs. The forward primers were the NBLs discussed in this paper. One could therefore located in exon 6 with the sequences: IF (50-GAAATCTCTTGC- speculate that a screwed expression of either UBE4B/ TAAACACTG-30) and IIF (50-TAATGGAAGTGCTAAT- UFD2 (located in 1p36.3) or UBE4A (in 11q23.3) is GATG-30) respectively. They were used in RT–PCR somehow an important step in NBL development. The amplification reactions together with a common reverse primer data prompt for further study of the role of UBE4B/ located in exon 8: R (50-TCCAACTCGGTCGAAACACTC-30) UFD2, for example, in experimental animal models and such studies will be performed.

were preferentially of maternal origin suggesting pater- nal imprinting of the locus. Several LOH studies including a large number of NBL tumors have deﬁned Acknowledegements We gratefully acknowledge the ﬁnancial support of the 1p36.3 as the region to be consistently deleted with Swedish Cancer Society, the Children’s Cancer Foundation, varying SROs (Ichimiya et al., 1999; Bauer et al., 2001; the IngaBritt and Arne Lundberg Foundation, the Assar Maris et al., 2001; White et al., 2001). The absence of Gabrielsson Foundation, the Wilhelm and Martina Lundgren mutations in the coding regions of the investigated genes Science Foundation and the Sahlgrenska University Hospital is consistent with the observation by Caron et al. (1995, Foundation.

Oncogene Neuroblastoma tumor suppressor candidate region C Krona et al 2351 References

Abel F, Ejeska¨ r K, Kogner P and Martinsson T. (1999). Br. J. Lal S, Choi JH, Shaw JR and Hannah LC. (1999). Plant. Cancer, 81, 1402–1409. Physiol., 121, 411–418. Abel F, Sjo¨ berg RM, Ejeska¨ r K, Krona C and Martinsson T. Lastowska M, Cotterill S, Pearson AD, Roberts P, McGuckin (2002). Br. J. Cancer, 86, 596–604. A, Lewis I and Bown N. (1997). Eur. J. Cancer, 33, 1627– Antonarakis SE. (1998). Hum. Mutat., 11, 1–3. 1633. Bauer A, Savelyeva L, Claas A, Praml C, Berthold F and Liu X, Li P, Widlak P, Zou H, Luo X, Garrard WT and Wang Schwab M. (2001). Genes Chromosomes Cancer, 31, 228–239. X. (1998). Proc. Natl. Acad. Sci. USA, 95, 8461–8466. Bown N, Cotterill S, Lastowska M, O’Neill S, Pearson AD, Liu X, Zou H, Slaughter C and Wang X. (1997). Cell, 89, 175– Plantaz D, Meddeb M, Danglot G, Brinkschmidt C, 184. Christiansen H, Laureys G and Speleman F. (1999). N. Maris JM, Guo C, Blake D, White PS, Hogarty MD, Engl. J. Med., 340, 1954–1961. Thompson PM, Rajalingam V, Gerbing R, Stram DO, Brodeur GM. (1990). Cancer Surv., 9, 673–688. Matthay KK, Seeger RC and Brodeur GM. (2001). Med. Brodeur GM, Pritchard J, Berthold F, Carlsen NL, Castel V, Pediatr. Oncol., 36, 32–36. Castelberry RP, De Bernardi B, Evans AE, Favrot M, Maris JM and Matthay KK. (1999). J. Clin. Oncol., 17, 2264– Hedborg F, Kaneko M, Kemshead J, Lampert F, Lee RE, 2279. Look AT, Pearson AD, Philip T, Roald B, Sawada T, Seeger Maris JM, White PS, Beltinger CP, Sulman EP, Castleberry RC, Tsuchida Y and Voute PA. (1993). J. Clin. Oncol., 11, RP, Shuster JJ, Look AT and Brodeur GM. (1995). Cancer 1466–1477. Res., 55, 4664–4669. Brodeur GM, Seeger RC, Barrett A, Berthold F, Castleberry Martinsson T, Sjo¨ berg RM, Hallstensson K, Nordling M, RP, D’Angio G, De Bernardi B, Evans AE, Favrot M, Hedborg F and Kogner P. (1997). Eur. J. Cancer, 33, 1997– Freeman AI, Haase G, Hartmann O, Hayes FA, Helson L, 2001. Kemshead J, Lampert F, Ninane J, Ohkawa H, Philip T, Martinsson T, Sjo¨ berg RM, Hedborg F and Kogner P. (1995). Pinkerton CR, Pritchard J, Sawada T, Siegel S, Smith EI, Cancer Res., 55, 5681–5686. Tsuchida Y and Voute PA. (1988). J. Clin. Oncol., 6, 1874–1881. Miller JR, Hocking AM, Brown JD and Moon RT. (1999). Burset M, Seledtsov IA and Solovyev VV. (2000). Nucleic Oncogene, 18, 7860–7872. Acids Res., 28, 4364–4375. Ohira M, Kageyama H, Mihara M, Furuta S, Machida T, Cadigan KM and Nusse R. (1997). Genes Dev., 11, 3286–3305. Shishikura T, Takayasu H, Islam A, Nakamura Y, Caron H, Peter M, van Sluis P, Speleman F, de Kraker J, Takahashi M, Tomioka N, Sakiyama S, Kaneko Y, Toyoda Laureys G, Michon J, Brugieres L, Voute PA, Westerveld A, A, Hattori M, Sakaki Y, Ohki M, Horii A, Soeda E, Slater R, Delattre O and Versteeg R. (1995). Hum. Mol. Inazawa J, Seki N, Kuma H, Nozawa I and Nakagawara A. Genet., 4, 535–539. (2000). Oncogene, 19, 4302–4307. Caron H, Spieker N, Godfried M, Veenstra M, van Sluis P, de Peifer M and Polakis P. (2000). Science, 287, 1606–1609. Kraker J, Voute P, and Versteeg R. (2001). Genes Chromo- Sakahira H, Enari M and Nagata S. (1999). J. Biol. Chem., somes Cancer, 30, 168–174. 274, 15740–15744. Clark F and Thanaraj TA. (2002). Hum. Mol. Genet., 11, 451– Sakahira H, Iwamatsu A and Nagata S. (2000). J. Biol. Chem., 464. 275, 8091–8096. Ejeska¨ r K, Abel F, Sjo¨ berg R, Backstrom J, Kogner P and Spieker N, Beitsma M, Van Sluis P, Chan A, Caron H and Martinsson T. (2000). Cytogenet. Cell Genet., 89, 62–66. Versteeg R. (2001). Genes Chromosomes Cancer, 31, 172– Ejeska¨ rK,Sjo¨ berg RM, Abel F, Kogner P, Ambros PF and 181. Martinsson T. (2001). Med. Pediatr. Oncol., 36, 61–66. Tago K, Nakamura T, Nishita M, Hyodo J, Nagai S, Murata Guo C, White PS, Weiss MJ, Hogarty MD, Thompson PM, Y, Adachi S, Ohwada S, Morishita Y, Shibuya H and Stram DO, Gerbing R, Matthay KK, Seeger RC, Brodeur Akiyama T. (2000). Genes Dev., 14, 1741–1749. GM and Maris JM. (1999). Oncogene, 18, 4948–4957. Thanaraj TA and Clark F. (2001). Nucleic Acids Res., 29, Guo C, White PS, Hogarty MD, Brodeur GM, Gerbing R, 2581–2593. Stram DO and Maris JM. (2000) Med. Pediatr. Oncol., 35, White PS, Maris JM, Beltinger C, Sulman E, Marshall HN, 544–546. Fujimori M, Kaufman BA, Biegel JA, Allen C, Hilliard C, Hochstrasser M. (1996). Annu. Rev. Genet., 30, 405–439. Valentine MB, Look AT, Enomoto H, Sakiyama S and Ichimiya S, Nimura Y, Kageyama H, Takada N, Sunahara M, Brodeur GM. (1995). Proc. Natl. Acad. Sci. USA, 92, 5520– Shishikura T, Nakamura Y, Sakiyama S, Seki N, Ohira M, 5524. Kaneko Y, McKeon F, Caput D and Nakagawara A. White PS, Thompson PM, Seifried BA, Sulman EP, Jensen SJ, (1999). Oncogene, 18, 1061–1066. Guo C, Maris JM, Hogarty MD, Allen C, Biegel JA, Matise Koegl M, Hoppe T, Schlenker S, Ulrich HD, Mayer TU and TC, Gregory SG, Reynolds CP and Brodeur GM. (2001). Jentsch S. (1999). Cell, 96, 635–644. Med. Pediatr. Oncol., 36, 37–41.

Oncogene