I I !4 ' 1996 Nature Publishing Group http://www.nature.com/naturegenetics
`
`letters
`
`Low incidence of BRCA2
`mutations in breast
`carcinoma and other cancers
`
`David H.-F. Teng’, Robert Bogden’,
`Jeffrey Mitchell’, Michelle Baumgard’,
`Russell Bell’, Simi Berry’, Thaylon Davis’,
`Phuong C. Hal, Robert Kehrer’,
`Srikanth Jammulapati’, Qian Chen’,
`Kenneth Offit 2, Mark H. Skolnick’,
`Sean V. Tavtigian’, Suresh Jhanwar2 ,
`Brad Swedlund’, Alexander K.C. Wong’ &
`Alexander Kamb’
`
`ty of BRCA2. Sixteen sequence-tagged sites (STSs)
`spanning about 650 kb of genomic DNA were used in
`PCR experiments to test for amplification from
`genorriic DNA (Fig. 1). Reproducible absence of
`amplification by particular primer pairs suggested
`homozygous deletion; the use of cell lines facilitated
`the analysis since primary specimens often contain
`some normal tissue that readily supports amplifica-
`tion 11 . We identified homozygous deletions in three
`cell lines - two colon (LS174 and SW620) and one
`breast (BT20) - by failure of a single STS (B52FT7)
`to amplify, use of several overlapping STSs confirmed
`this result. The deletion boundaries appeared identical
`in the three lines and were located roughly 6 kb apart.
`The deleted region lies nearly 350 kb 5’ of the BRCA2
`transcriptional start site and is separated from BRCA2
`by at least one other gene (S. Tavtigian, unpublished
`data), and is, thus, unlikely to interfere with BRCA2
`function. In addition, the deletion appears to be a
`common polymorphism based on analysis of normal
`DNAs using a deletion-specific primer pair as a probe
`(see Methods). The homozygous deletion detected in
`the cell lines was the result of either loss of heterozy-
`gosity (LOH) on one homolog, which uncovered the
`deletion (BT20 and LS174), or homozygosity of the
`polymorphic deleted chromosome (SW620).
`Having found no evidence for homozygous deletion
`of BRCA2
`in our experiments, we investigated
`whether smaller lesions such as point mutations con-
`tributed to BRCA2 inactivation. Since compound
`mutant heterozygotes and mutant homozygotes are
`rare, tumour suppressor gene inactivation nearly
`always involves LOH. The remaining allele, if inactive,
`typically contains disruptive mutations. Thus, if the
`goal is to identify mutations in tumour suppressor
`genes, it is sensible to preselect tumours or cell lines
`that exhibit LOH at the locus of interest 2 .
`We examined a group of 104 primary breast tumour
`samples and a set of 269 cell lines for LOH in the
`BRCA2 region. For the primary tumours, we quantita-
`tively compared amplifications of three short tandem
`repeat markers (STR4247, STR257, STR56IA) using
`fluorescence. Based on their combined heterozygosity
`indices, the chance that these markers will all be
`homozygous in a particular individual (assuming
`
`Inherited mutant alleles of familial tumour sup-
`pressor genes predispose individuals to particular
`types of cancer. In addition to an involvement in
`inherited susceptibility to cancer, these tumour
`suppressor genes are targets for somatic muta-
`tions in sporadic cancers of the same type found
`in the familial forms’. An exception is BRCA1,
`which contributes to a significant fraction of famil-
`ial breast and ovarian cancer, but undergoes
`mutation at very low rates in sporadic breast and
`ovarian cancers 21. This finding suggests that
`other genes may be the principal targets for
`somatic mutation in breast carcinoma. A second,
`recently identified familial breast cancer gene,
`BRCA2 (refs 5-8), accounts for a proportion of
`breast cancer roughly equal to BRCA1. Like
`BRCA1, BRCA2 behaves as a dominantly inherit-
`ed tumour suppressor gene. Individuals who
`inherit one mutant allele are at increased risk for
`breast cancer, and the tumours they develop lose
`the wild-type allele by heterozygous deletion 9 .
`The BRCA2 coding sequence is huge, composed
`of 26 exons that span 10,443 bp 8. Here we investi-
`gate the rate of BRCA2 mutation in sporadic
`breast cancers and in a set of cell lines that repre-
`sent twelve other tumour types. Surprisingly,
`mutations in BRCA2 are infrequent in cancers
`including breast carcinoma.
`However, a probable germline
`100 kb
`mutation in a pancreatic tumour -
`cell line suggests a role for
`BRCA2 in susceptibility to pan-
`creatic cancer.
`Tumour suppressor genes can be
`inactivated by homozygous dele-
`tion; indeed, detection of a
`homozygous deletion in a pancreat-
`ic xenograft was instrumental in the Fig.
`Schulte at at. deletion
`I Screen for ho
`nozygous deletions in tumour cell lines. The positions of STS mark-
`effort to isolate BRCA2 by position- ers used in the analy
`sis are shown. D13S260 and 013S165, SIR markers that help on-
`al cloning’0. The chance observa- ant the physical an
`id the genetic maps, are underlined. The three SIRs used to
`tion of a BRCA2 homozygous determine [OH are b
`oxed. Sequences of the STS primers are available upon request. In
`genes within this region, several cDNA fragments were lined
`deletion suggested that other events an earlier search fo
`together into a cand
`idate gene (CG029) that includes sequences within the deletion (S.
`might occur in breast tumours or in Tavtigian, unpublish
`id data). No splice junctions were identified within CG029, but a
`edominantly a 2.4-kb mRNA species from testis and heart on a
`additional cancer types. We probe recognized P1
`poly(A) northern bl
`ot. The locations and transcription orientations of RRCA2 and
`screened a set of 150 cell lines
`CG029 are shown b
`y arrows. The deletion polymorphism location is shown by a trian-
`derived from different cancers for gle. Also shown is
`he extent of the homozygous deletion identified in a pancreatic
`ntromere; TEL, telorilere.
`10 . CEN, ce
`homozygous deletions in the vicini- xenograft
`
`, (cid:9)
`
`l
`
`I (cid:9)
`
`II (cid:9)
`
`I (cid:9)
`
`I (cid:9)
`
`I (cid:9)
`
`I (cid:9)
`
`I (cid:9)
`
`II (cid:9)
`
`I (cid:9)
`
`II (cid:9)
`
`I (cid:9)
`
`III (cid:9)
`
`I (cid:9)
`
`I (cid:9)
`
`I
`
`CEN (cid:9)
`
`C0029 (cid:9)
`
`8RCA2 (cid:9)
`
`I (cid:9)
`
`I
`
`TEL
`
`5Myriad Genetics,
`Inc., Wakara Way,
`Salt Lake City, Utah
`84108, USA
`2Department of
`Human Genetics,
`Memorial Sloan-
`Kettering Cancer
`Center, New York,
`New York 10021,
`USA
`
`Correspondence
`should be addressed
`to AK.
`e-mail: akamb@
`myriad.com
`
`nature genetics volume 13 june 1996 (cid:9)
`
`241
`
`GeneDX 1019, pg. 1
`
`

`

`letters (cid:9)
`
`a !4 ' 1996 Nature Publishing Group http://www.nature.com/naturegenetics
`
`
`
`Fig. 2 LOH analysis of prima-
`ry breast tumours. Alleles of
`STR markers are indicated
`below the chromatogram. a,
`Example of a tumour het-
`erozygous at BFICA2. b,
`Example of a tumour with
`LOH at BRCA2. Fluores-
`cence units are on the ordi-
`nate; size in basepaira is on
`the abcissa. T, tumour; N,
`normal.
`
`280 (cid:9)
`
`340
`
`260 (cid:9)
`
`340
`
`4800
`
`2400
`
`5600
`
`MiN
`
`260 (cid:9)
`I (cid:9)
`
`I
`
`340
`
`b (cid:9)
`
`5600
`
`2800 (cid:9)
`
`N
`
`1
`
`260 (cid:9)
`I (cid:9)
`
`I (cid:9)
`
`T
`
`24M A
`
`340
`I
`
`I (cid:9)
`
`H
`
`linkage equilibrium) is only one in 250. Even if there is
`LOH, the presence of normal cells in the primary
`tumour sample makes it impossible to entirely elimi-
`nate signal from the lost allele. Rather, the relative
`intensities of the two alleles are altered as can clearly be
`seen by comparing the allelic peak heights from nor-
`mal tissue with peak heights from the tumour (Fig. 2).
`Using this analysis, we classified 30 tumours (29%) as
`having LO1-I at the BRCA2 locus (Table 1); this figure
`is similar to previous estimates 92 .
`We assessed LOH in the cell lines in a different fash-
`ion. Since homozygosity of all three STRs together in a
`single sample was improbable, and since normal cells
`were not present, we interpreted apparent homozygosi-
`ty at all STRs as LOH in the BRCA2 region. Using this
`criterion, 851269 of the cell lines exhibited LOM (Table
`1). The frequencies in lines of different tumour types
`varied: 4/6 ovarian cell lines and 31/62 lung cancer
`lines displayed LOH compared with 17/81 melanoma
`
`st tumours (cid:9)
`Number (cid:9)
`sequenced (cid:9)
`
`lines and 2/11 breast cancer lines.
`We screened the group of 30 primary breast cancers,
`preselected for LOH in the BRCA2 region (Table 1), by
`DNA sequence analysis for sequence variants. We
`examined greater than 95% of the coding sequence
`and splice junctions of BRCA2 in each sample. One
`specimen contained a frameshift mutation, one a mis-
`sense and a frameshift mutation, one a nonsense
`mutation, and one a missense mutation (Table 2a).
`The nonsense mutation would delete 156 codons at
`the C terminus, suggesting that the C-terminal end of
`BRCA2 is important for tumour suppressor activity.
`All sequence variants were also present in the corre-
`sponding normal DNA from these cancer patients
`indicating that these were germline lesions. To exclude
`the unlikely possibility that preselection for LOH
`introduced a systematic bias against detecting muta-
`tions (for example, dominant mutations or compound
`heterozygotes), 12 samples heterozygous at BRCA2
`were also screened. Three of these
`revealed missense changes that were
`also found in the corresponding
`normal samples. Thus, in a set of 42
`breast carcinoma samples, 30 of
`which displayed LOH at the BRCA2
`locus, no somatic mutations were
`identified. The frameshift and non-
`sense changes are likely to be pre-
`mutations (cid:9)
`that
`disposing (cid:9)
`influenced development of breast
`cancer in these patients. The mis-
`sense variants are rare; they were
`each observed only once during
`analysis of 115 chromosomes. From
`these data, however, it is not possi-
`ble to distinguish between rare neu-
`and
`polymorphisms (cid:9)
`tral (cid:9)
`predisposing mutations.
`
`Table 1 LOH analysis of cell lines and primary brea
`Percentage
`Number
`LOH /screened
`Type
`LOH
`Astrocytoma
`32%
`6/19
`Bladder
`35%
`6/17
`Breast
`18%
`2/11
`25%
`2/8
`Colon
`Glioma
`31%
`11/36
`Lung
`50%
`31/62
`0
`Lymphoma
`0/4
`Melanoma
`21%
`17/81
`10%
`1/10
`Neuroblastoma
`67%
`Ovarian
`4/6
`Pancreatic
`33%
`1/3
`Prostate
`0
`0/2
`Renal
`40%
`4/10
`33% (avg. = 21
`Total
`3%) (cid:9)
`85/269
`42 (cid:9)
`29%
`Primary breast
`30/104
`Percentage L011 was calculated two ways: as total and as a mean of percentages (avg.).
`
`2 (cid:9)
`2 (cid:9)
`5 (cid:9)
`20 (cid:9)
`
`1 (cid:9)
`4 (cid:9)
`1 (cid:9)
`
`58 (cid:9)
`
`242 (cid:9)
`
`nature genetics volume 13 june 1996
`
`GeneDX 1019, pg. 2
`
`(cid:9)
`(cid:9)
`

`

`' 1996 Nature Publishing Group http://www.nature.com/naturegenetics
`
`
`
`letters
`
`Table 2a Mutations identified in BRCA2
`Germline
`Change
`Effect
`Type
`Sample
`LOH
`G451
`Ala-9Pro
`nd.
`Renal
`41-15
`yes
`Al 093C
`Asrt-+His
`nd.
`Ovarian
`4G1
`yes
`G1291C
`Val-*Leu
`n.d.
`Lung
`2F8
`yes
`1493delA
`frameshift
`Primary breast
`BT1 10
`yes
`yes
`C2117T
`Thr-Ale
`Ovarian
`4F8
`nd.
`yes
`A2411C
`Primary breast
`BT163
`yes
`no
`Asp-+Ala
`G4813A
`Gly-sArg
`Bladder
`1D6
`n.d.
`no
`T5868G
`Asn-sL.ys
`Primary breast
`81333
`yes
`no
`61 74delT
`frameshift
`Primary breast
`BT1 11
`yes
`yes
`4G3
`yes
`as above
`as above
`Pancreatic
`n.d.
`187
`C6328T
`Arg-4Cys
`Astrocytoma
`n.d.
`yes
`G7049T
`Primary breast
`BT1 18
`no
`GIy-sVal
`yes
`G7491C
`BT1 15
`yes
`G!n-*His
`Primary breast
`yes
`A9537G
`Melanoma
`3D5
`nd.
`yes
`lle-#Met
`Al 02041
`Lys-sstop
`Primary breast
`8185
`yes
`yes
`1E4
`Breast
`nd.
`yes
`C10298G
`Thr-9Arg
`Al 0462G
`Primary breast
`BT1 10
`lle-sVal
`yes
`yes
`Listed are the mutation positions based on the Genbank entry of BRCA2 (Accession
`number HSU43746)8. As above means that the tumour has the same change as the
`sample listed before it. nd., not determined.
`
`-
`
`Table 2b Common polymorphisms and silent
`substitutions detected in BRCA2 by DNA sequencing
`Position
`Change
`Effect
`Frequency
`5UTR(203)
`G/A
`0.32(0.26)
`Asn-+l-lis
`PM(1 342)
`NC
`0.32(0.37)
`PM(2457)
`silent
`TIC
`004 (0.05)
`PM(3199)
`Asn-9Asp
`0.04(0.08)
`NG
`PM(3624)
`silent
`NG
`0.35
`AIG
`PM(3668)
`Asn-sSer
`0(0.15)
`PM(4035)
`1/C
`silent
`0.24(0.10)
`PM(5972)
`Thr-sMet
`C/T
`0.03
`PM(7470)
`silent
`0.26(0.15)
`NG
`1593
`silent
`A--,)G
`<0.01
`4296
`0-+A
`silent
`<0.01
`5691
`A--)G
`silent
`<0.01
`6051
`silent
`A--9G
`<0.01
`6828
`silent
`T-aC
`<001
`6921
`T---)C
`silent
`<0.01
`Since some rare silent variants may affect gene function (for
`example, splicing)21 , these are not preceded by ’PM". The fre
`quencies of polymorphisms shown involve the second of the
`nucleotide pair. Frequencies reported in a previous study are
`shown in parentheses8. Numbering is as in Table 2a.
`
`The technical ease with which sequence changes can
`be detected in cell lines, and the opportunity to exam-
`ine BRCA2 in cell lines derived from tumours other
`than breast, led us to screen a group of 58 cell lines
`preselected for LOH at BRCA2 (Table 1). We screened
`greater than 95% of the BRCA2 coding sequence of
`each sample, but only one clearly disruptive mutation,
`a frameshift, was identified by DNA sequence analysis
`(Table 2a). This mutation (6174delT) was present in a
`pancreatic cancer line and is identical to one found in
`the germline of primary tumour, BT1 11, and to previ-
`ously detected germline frameshift lesions" , ". These
`findings suggest that this particular frameshift may be
`a relatively common germline BRCA2 mutation. In
`addition, eight missense sequence variants were
`detected (Table 2a); however, it was not possible to
`determine whether these sequence variants were
`somatic and whether they compromise BRCA2 protein
`function.
`Detection of a probable germline BRCA2 mutation
`in a pancreatic tumour cell line suggests that BRCA2
`mutations may predispose to pancreatic cancer. This
`mutation also adds weight to the involvement of
`BRCA2 in sporadic pancreatic cancer, implied previ-
`ously by the homozygous deletion observed in a pan-
`creatic xenograftt 0. Since we examined only three
`pancreatic cell lines, further investigation of BRCA2
`mutations in pancreatic cancers is warranted.
`The lack of evidence for somatic inactivation places
`BRCA2 in a select company of familial tumour sup-
`pressor genes that are infrequently mutated in spo-
`radic tumours. Among tumour suppressor genes
`identified through kindred studies so far, only BRCA1,
`WTI and possibly hMLH1 and hMSH2 exhibit such
`features. It is conceivable that modes of inactivation
`other than deletion and mutation, such as methyla-
`tion-silencing, may occur 14 ’ 15. However, in no case has
`methylation been identified as the sole mechanism for
`inactivation of a specific tumour suppressor gene. It is
`also unlikely that a large number of mutations were
`missed since we screened nearly the entire coding
`sequence in every sample and detected polymorphisms
`at frequencies consistent with previous work (Table
`
`2b). In addition, most germline BRCA2 mutations
`described so far are of the frameshift variety 7’8 . These
`are especially easy to spot. Thus, it is probable that
`BRCA2 is not inactivated at appreciable frequencies in
`sporadic breast tumours, despite its prominent role in
`hereditary breast cancer.
`The observation, however, that roughly a third of
`breast tumours and tumour cell lines have lost one
`copy of BRCA2 may suggest a role for another tumour
`suppressor located nearby on chromosome 13. The
`retinoblastoma locus, RBI, lies on the short arm of
`chromosome 13 and may be responsible for some
`LOH, particularly in lung cancer lines’ 6 " 7. In addition,
`random chromosome deletion may contribute to LOH
`in the region as described for other chromosomal loca-
`tions 18’ 19 . As was the case with BRCAJ, attempts to use
`LOH in sporadic breast tumours to localize BRCA2
`would have given false information 20 .
`The hope that study of familial cancer genes will
`yield insight into tumour progression has been amply
`justified. In the case of breast carcinoma, the spectrum
`of genes influencing predisposition through germline
`mutation may be different from the genes inactivated
`by somatic mutation during progression. Nevertheless,
`studies of BRCAI and BRCA2 provide the opportunity
`to gain fundamental insights into growth control path-
`ways that operate within breast epithelial cells.
`
`Methods
`DNA preparation. Primary breast tumour and normal sample
`pairs were anonymously obtained from the Memorial Sloan-
`Kettering Cancer Center. Patient histories are not available, but
`the sample is random, unselected for age of onset or family his-
`tory. Cancer cell lines were purchased from the ATCC. DNA
`from cell lines was prepared as described’ 9. DNA from tissue
`was isolated using the Easy-DNA kit (Invitrogen).
`
`PCR amplification. A total of 33 arnplicons were generated by
`PCR that encompass the 26 coding exons of BRCA2. Nested
`primer sets were used for all amplifications. In general, the
`PCR conditions were: an initial single denaturation step at
`95 (cid:176)C for 1 min followed by cycles of denaturing at 95
`)C (6 s),
`annealing at 55 (cid:176)C (15 s) and extension at 72(cid:176)C (I mm). For
`the primary amplification step, 1 to 10 ng of genonlic DNA
`
`nature genetics volume 13 iune 1996 (cid:9)
`
`243
`
`GeneDX 1019, pg. 3
`
`

`

`letters
`
`I
`
`(cid:149) 54
`
`' 1996 Nature Publishing Group http://www.nature.com/naturegenetics
`
`were subjected to a 20 cycle reaction after which the primary
`PCR products were diluted ten-fold and reamplifled using
`nested primers for another 33 cycles.
`
`DNA sequencing and mutation screening. DNA sequencing
`was carried out either on the ABI 377 (Applied Biosystems
`Division, Perkin-Elmer) or manually. For the radioactive
`mutation screen, the amplified products were purified on
`agarose gels followed by Qiaquick (Qiagen). DNA sequence
`was generated using the Cyclist sequencing kit (Stratagene)
`and resolved on 6% polyacrylamide gels. In parallel, non-
`radioactive sequencing using fluorescent labelling dyes was
`performed using the TaqFS sequencing kit followed by elec-
`trophoresis on ABI 377 sequencers. Samples were gridded into
`96-well trays to facilitate PCR and sequencing. Dropouts of
`particular PCR and sequencing reactions were repeated until
`>95% coverage was obtained for every sample. Sequence data
`were analysed with the Sequencher software (Genes Codes). All
`detected mutations were confirmed by sequencing a newly
`amplified PCR product to exclude the possibility of PCR arte-
`facts. Primers for PCR and sequencing were in roughly half the
`eases identical to those used previously and are available upon
`request8 .
`
`Loss of heterozygosity analysis. For genotyping, -10 ng of
`genomic DNA was amplified by PCR with the following three
`sets of fluorescently tagged STRs (5’.->3’):
`(1) mM4247.4A.2F1 ACCATCAAACACATCATCC
`mM4247.4A.2R2 AGAAAGTAACTTGGAGGGAG
`
`(2)
`
`CTCCTGAAACTGTTCCCTTGG
`STR257-FC (cid:9)
`TAATGGTGCTGGGATATTTGG
`STR257-RD (cid:9)
`(3) mMB56IA-3. 1FA2 GAATGTCGAAGAGCTTGTC
`mMB56IA-3. 1RB AAACATACGCTTAGCCAGAC
`
`quantified with Genescan software. For tumours, clear peak
`height differences between alleles amplified from normal and
`tumour samples were scored as having LOH. For cell lines, if
`one STR was heterozygous, the sample was scored as non-
`LOl-1. In only one case was a cell line or tumour miscalled
`based on later analysis of single base polymorphisms. The het-
`erozygosity indices for the markers are: STR4247 = 0.89;
`STR257 = 0.72; STR561A = 0.88 (S. Neuhausen, personal
`communication; B. Swedlund, unpublished data).
`
`Analysis of deletion polymorphism. A deletion-specific
`primer pair and a wild type-specific pair was used to assess the
`frequency of the polymorphism in a set of 87 control germline
`DNA samples. Only one sample appeared homozygous, where-
`as seven were heterozygous; thus, the overall deletion frequen-
`cy is 5% of chromosomes. The primer pairs used were
`R130-A19 which generates a 3-kb product from wild-type
`genomic DNA, and R115-A19 which amplifies across the delet-
`ed region to yield a 2-kb fragment. The primer sequences were
`(5-sd’):
`
`4353.R 115 GGC ACC TAT ATC CAC AGA CA
`4353.R 130 AAG ACT CCT GGA GTC TAG AAC
`4353.A19 CCA AAG ATT TAG TGT AAG CAG AAC
`
`GenBank accession number. BRCA2: HSU43 746.
`
`Acknowledgements
`We thank P Barrelfor comments on the manuscript, S.
`Neuhausen for providing heterozygosity index values, and I.
`Collett and T. Tran for technical assistance.
`
`The PCR products were resolved using the AB1377 and
`
`Received 19 March; accepted 24 April 1996.
`
`Knudson, A. C. Al a the (cancer) family. Nature Genet. 5, 103-104(1993).
`2. Futreal, P. A. at al. BRCAI mutations in primary breast and ovarian
`carcinomas. Science 266,120-122 (1994).
`3. Hosting, L. eta). A somatic 8RCAI mutation in an ovarian tumor. Nature
`Garret. 9,343-344(1995).
`4. Merajver. 3, D. at al. Somatic mutations in the SRCAI gene in sporadic
`ovarian tumor Nature Garret. 9.439-443(1995).
`5. Wooster, R. et a/. Localization of a breast cancer susceptibility gene.
`BRCA2, to chromosome 13q1 2-13. Science 265,2088-2090(1994).
`6. Thortacius, S. at al. Linkage to BRCA2 region in hereditary male breast
`cancer. Lancet 346,544-545)1995).
`7. Wooster, H. et al. Identification of the breast cancer susceptibility gene
`BRCA2. Nature 378,789-792(1995).
`8. Tavtigian, S. V. at al. The complete BRCA2 gene and mutations in
`chromosome 13q-linked kindreds. Nature Garret. 12. 1-6(1996).
`9. Collins, N. at al. Consistent loss of the wild type allele in breast cancers
`from a family linked to the BHCA2 gene on chromosome 13q12.-13.
`Oncogene ID, 1673-1675(1995).
`10. Schutte, M. at a!, Identification by representational difference analysis of a
`homozygous deletion in pancreatic carcinoma that lies within the BRCA2
`region. Proc. Nat). Aced. Sd. USA 92,5950-5954)1996).
`11. Kamb, A. at al. Response to Rates of p16 (MTSI) Mutations in Primary
`Tumors with 9p Loss. Cairns at al. 1995 Science 265,416-417(1994).
`12. Cleton-Jansen, A. M. at al. Loss of troterozygosity in sporadic breast
`tumours at the BFICA2 locus on chromosome 13q12-q13. Be J. Cancer
`
`72,1241-1243(1995).
`13. Neuhausen, S. etal. Recurrent BRCA2 6174deiT mutations in Ashkenazi
`Jewish women affected by breast cancer. Nature Garret. 13, 126-128
`(1996).
`14. Gonzalez-Zulueta, M. at al. Methylatiori of the 5’ CpG island of the
`p161CDKN2 tumor suppressor gene in normal and a transformed
`human tissues correlates with gene silencing. CancerRes. 55,4531-4535
`(1995).
`15, Mario, k at al. 5’ CpG island methylation is associated with transcriptional
`silencing of the tumour suppressorpl6lCDKN2/MTSI in human cancers.
`NatumMed. 1,686-691(1995).
`16. Yokota. J. at at. Altered expression of the retinoblaatocna (RB) gene in
`smat-cell carcinoma of the lung. Oncogene 3,471-675(1988).
`17. Hansel, C. H. at al. Altered structure and expression of the human
`retinoblastoma susceptibility gene in small cell lung cancer. Cancer Pea
`50,3067-3072(1990).
`18. Vogelstein, S. at al. Allelotype of colorectal carcinomas. Science 244,
`207-211)1989).
`19. Liu. Q. etal. CDKN2 (MTS1 (tumor suppressor gene mutations in human
`tumor call lines. Oncogerre 10, 1061-1067(1995).
`20. Cmi, M. at at A novel rnetalloprotease/disintegnn-like gene at 17q21.3 is
`somatically rearranged in two primary breast cancers, Nature Garret. 5,
`151-157(1993).
`21 . Richard, I. & Beckmann, J. S. How neutral are synonymous codon
`mutations? Nature Genet. 10, 259 (1995).
`
`244 (cid:9)
`
`nature genetics volume 13 june 1996
`
`GeneDX 1019, pg. 4
`
`(cid:9)
`