- Email: [email protected]
Exclusion of the familial melanoma locus (MLM) from the PNDD1S47 and MYCL1 regions of chromosome arm 1p in 7 Australian pedigrees
GENOMICS
12,
18-25
(19%)
Exclusion of the Familial Melanoma Locus (MLM) from the PND/ DlS47 and MYCLI Regions of Chromosome Arm Ip in 7 Australian Pedigrees DEREK J. NANCARROW, JANE M. PALMER, MARILYN K. WALTERS, BEVERLEY M. KERR, GREGORY J. HAFNER, LUKE GARSKE, G. RODERICK MCLEOD,* AND NICHOLAS K. HAYWARD’ Queensland Cancer Fund Research Unit, Joint Experimental Oncology Program, Queensland Institute of Medical Research, Herston, 4029, Australia; and *Queensland Melanoma Project, Princess Alexandra Hospital, Wooloongabba, 4 702, Australia ReceivedJune
17. 1991;revisedAugust23,1991
velopment Conference, 1984), in which multiple, atypical moles are frequently found in association with malignant melanoma. Thus DNS formed the physical marker for familial melanoma carriers based on this circumstantial evidence. Using the combination of melanoma and DNS as the disease phenotype, Greene et al. (1983) found a suggestion of loose linkage (Z,, = 2 at 0 = 0.3) in 14 dysplastic nevus syndrome/familial melanoma (DNS/MLM) families to the rhesus blood group locus on the short arm of chromosome one. Four years later Bale et al. (1987) verified this linkage to rhesus and several other markers on lp. In 1989 the same group claimed to have localized this putative DNS/MLM gene, designated CMM, to a region between the anonymous DNA marker DlS47 and the gene for pronatriodilatin (PND), 36 CM (Dracopoli et aZ., 1991) telomerit to rhesus, in six families (Bale et al., 1989). Coinciding with this report, van Haeringen et al. (1989) published data that claimed to exclude all of the short arm of chromosome 1 in six Dutch DNS or DNS/MLM families, and in 1990, the same group followed up these findings (in the same six families) with a more specific exclusion, using the DlS47 and PND probes (Gruis et al., 1990). Two other groups have published data excluding the DlS47/PND region from linkage to CMM (DNS/ MLM) using the same model for DNS/MLM as Bale et al. (1989). Cannon- Albright et al. (1990) presented exclusion data for this region in three Utah DNS/ MLM pedigrees, while Kefford et al. (1991) excluded this region in eight Australian kindreds. Both of these groups openly questioned combining MLM and DNS to study linkage and each paper included analyses of MLM alone. Cannon-Albright et al. (1990) discussed the subjective definition of DNS described by Greene et al. (1983) and the subjective nature of the diagnosis. Several of the Australian families presented by Kefford et al. (1991) showed little or no history of dys-
Familial melanoma (MLM) is sometimes found associated with the dysplastic nevus syndrome (DNS). Considerable controversy exists over the possible assignment of a cutane-
ous malignant melanomafdysplastic nevus gene, designated CMM, to the distal short arm of chromosome 1, linked to the PND and DlS47 loci. To date, no support for linkage of MLM alone to these markers has been found; likewise no study has been able to exclude the entire region between PND and DlS47 from linkage to MLM. We have carried out linkage studies between markers on lp and MLM in seven Australian kindreds; three of these are the largest reported worldwide. We have been able to exclude localization of an MLM gene from a 40-CM region that spans the interval between DlS47 and PND and extends approximately 15 CM on either side of these markers. In addition, we can exclude a region of about 20 CM around the MYCLl/DlSB? loci. Q 1992 Academic press, IX.
INTRODUCTION
When analyzing a genetic disease it is useful to have a physical marker that appears to be associated with the disease in question, a precursor, or an early symptom. This is particularly important for identifying family members who are carriers in familial diseasesthat are codominant, have late or variable age of onset, or have a reduced penetrance. Familial melanoma, although apparently dominant, has both an age-relatedonset and a reducedpenetrance, asdemonstrated by skipped generations in pedigrees worldwide. Lynch et al. (1978) and Clark et al. (1978) independently characterized what was later to be called the dysplastic nevus syndrome (DNS) (Consensus De1 To whom correspondence should be addressed Institute of Medical Research, Herston Rd., Queensland, Australia.
ossa-7543/92 $3.00 Copyright 0 1992 by Academic Press, All rights of reproduction in any form
at Queensland Herston, 4029,
18 Inc.
reserved.
EXCLUSION
U-IO-24
25-5-19 167-15 I
OF
FAMILIAL
MELANOMA
361-22 21-l-15 I
LOCUS
n-9-u
w3
FROM
15-Z-39 30-3-70 I
M-I-30 u-11-80 I
026
V.8
19
lp.
031
an
036
wn
FIG. 1. Pedigrees of the seven familial melanoma kindreds used in this study. (m) Affected male; (0) unaffected female; (EJ) deceased unaffected male; (h) deceased affected female. Only the numbered individuals were available for genotyping in this study. Alleles for the DlS47 locus and BglI/XhoI haplotypes of the PND locus respectively are given below each family member. Dates, in descending position above and left of each person, are date of birth, date of first melanoma, and date of death.
plastic nevus syndrome or surgical removal of histologically categorized dysplastic nevi. This led them to introduce a linkage model for melanoma alone, based on data from the Sydney Melanoma Unit. This model takes into consideration age of onset, sporadic mela-
noma, and incomplete penetrance (factors all seen in familial melanoma kindreds but usually left out of previous models). We present here a comprehensive analysis of the p32-p36 region of chromosome 1, carried out on six
20
NANCARROW
8/D a/a 014
B/D r/b 004
YY */a 003
B/6 r/o 042
ET
AL.
WE alo 043
m9
oaYm5mD FIG.
DO1
l-Continued
familial melanoma pedigrees collected in Queensland and one in Western Australia, three of which contain more living affecteds than any other kindred thus far presented. We have chosen to present our data using the model of Kefford et al. (1991) for melanoma alone (O-15 years 1% penetrance, 0% sporadic; 16-25 years 30% penetrance, 1% sporadic; 26-45 years 50% penetrance, 3% sporadic; 46 years and over 80% penetrance, 5% sporadic; with a gene frequency of 0.003) for several reasons; our close geographical proximity to families collected by the Kefford group, because we too have non-DNS families and because the data used to calculate parameters of the model were estimated
from a population based survey. We have, however, included an analysis with the model of Bale et al. (1989) (99% penetrance for all people over 20 years with a gene frequency of 0.001 and no allowance for sporadic melanoma) strictly for comparison with existing reports. The data presented here are the first to exclude the entire region between the PND and DlS47 loci from linkage to MLM alone, with both these models. METHODS DNA from 40 patients with histologically confirmed melanoma and 57 unaffected members of 7 fa-
EXCLUSION
TABLE
OF
FAMILIAL
1
Allele Sizes and Frequencies for Probes Used to Study Possible Linkage of Familial Melanoma to Loci on lp Probe CRI-336 pJAll0
LOCUS
DlS47 PND
Enzyme RSUI
xhd BglI
pLBK8B/E5
ALPL
BClI
pYNZ2
DlS57
RSd
PLmYc
MYCLl
EcoRI
pl-11B
DlS15
BglII
Allele size (kb)
Allele frequency
VNTR 12 10 10 6 7.4 4.3, 3.1 2.5
0.1666” 0.46 0.54 0.87 0.13 0.67 0.33 0.3334 0.3333 0.3333 0.44 0.56 0.56 0.30 0.14
1.8
1.5 10 6.6 19 17 14.8
Reference (9)
(21) (11)
6’5) POP (11) (11)
“To enable appropriate analysis with LIPED we have nominated six alleles of equal frequency. b Since we only observed three of the five alleles seen by Nakamura et al. (20) we have adjusted the frequencies accordingly.
milial melanoma pedigrees (Fig. 1) was obtained from EBV-transformed lymphocytes using the method of Miller et al. (1988). Only family 60001 has been published previously, in part (family B in Ramsay et al. (1982)). The other six families were obtained through the Queensland Melanoma Project (Princess Alexandra Hospital, Brisbane) and were selected by establishing whether probands had at least two other family members with melanoma, still living. Thus, in our study, MLM is defined as three or more living first- or second-degree relatives with histologically confirmed melanoma. In three families (40575, 40354, and 40750) there were no cases (either melanoma patients or unaffected relatives) of classical dysplastic nevus syndrome (Lynch et aZ., 1978; Clark et al., 1978) or isolated excision of a single (sporadic) nevus histopathologically categorized as dysplastic. Only one member (213) of family 40582 and two members (009, 021) of family 40599 had classical DNS. Each of these patients also had melanoma. A nurse or general practitioner asked all family members whether they had at any time had any pigmented lesions surgically removed. Histology reports were obtained for each of these lesions and the diagnosis noted. Histologically classified dysplastic nevi were excised from patients 212 in family 40582; 007 and 005 (two) in family 41001; and 002, 005, 017 (two), and 021 (four) in family 60001. Once again, all instances of surgical removal of DN were in family members who also had histopathologically confirmed melanoma. In contrast, there were no family members
MELANOMA
LOCUS
FROM
lp.
21
without melanoma who had DNS or had surgical excision of an isolated DN. All nevi excised from nonaffected family members (and some of the other nevi removed from affected cases) were of the benign compound, intradermal, or junctional type. For each RFLP studied, DNA from all 97 individuals was cleaved with the appropriate restriction enzyme, electrophoresed in 0.8-1.2% agarose gels (depending on the expected allele sizes), then capillary blotted onto nylon membranes using 10X SSC. Prehybridization, hybridization, and stringency washes were carried out according to the method of Church and Gilbert (1984). Probe DNA was labeled by the random priming method of Feinberg and Vogelstein (1983, 1984). The probes used in this study were chosen to correspond with those used in previously reported linkage analyses between MLM or CMM (DNS/MLM) and chromosome lp markers, and are listed in Table 1, along with their locus name, the enzyme used to generate the RFLP, and the corresponding allele frequencies. The latter were all derived from Caucasian populations (see Table 1 for references) and should thus be appropriate for our pedigrees, which are all AngloSaxon in origin. In the case of the PND locus two enzyme systems were characterised, XhoI and BgZI, then haplotyped for linkage analysis. The hypervariability of DlS47 and the inability to uniquely characterize alleles across families combined with the limitation of computer analysis to a maximum of six alleles at any one locus, resulted in the nomination of six alleles of equal frequency in the analysis of this locus; four for first generation alleles and the remaining two encompassing all spouse alleles. For DlS57, we only observed three alleles, where previously five had been reported (Nakamura et al., 1987), and gave these alleles equal frequencies. All linkage analyses were performed with the computer program LIPED (Ott, 1976), with TABLOID2 compiling these data directly to form Table 2. Two models were used to represent the mode of inheritance of familial melanoma; that of Kefford et al. (1991) and that of Bale et al. (1989). We chose not to use the only other model for MLM/DNS so far proposed (Gruis et cd., 1990) since it does not incorporate variable age of onset, reduced penetrances, or sporadic incidence of disease. All genetic distances discussed in this paper are those of sex average taken from the latest CEPH Consortium Linkage Map (Table 4 in Dracopoli et al. (1991)). RESULTS
Each of the 97 DNA samples (numbered individuals) obtained from the seven familial melanoma kindreds illustrated in Fig. 1 was characterized for the ’ A set of computer to compress LIPED
programs developed in this laboratory output; available upon request.
used
22
NANCARROW
TABLE Lod Scores for Linkage Analyzed with the Kefford and 0.4 Kefford Family
.oooo
.0500 Locus:
41001 60001 40599 40750 40582 40575 40354 Total
-2.184 -7.197 -2.152 ,140 -.005 ,000 -.017 -12.015
-.913 -2.895 -1.500 .109 .002 .ooo -.016 -5.213 Locus:
41001 60001 40599 40750 40582 40575 40354 Total
-1.275 -4.634 -3.244 -.013 -.027 .ooo -.012 -9.205
-.852 -2.399 -1.223 -.OlO -.025 .ooo -.008 -4.517 Locus:
41001 60001 40599 40750 40582 40575 40354 Total
.366 -.602 -2.153 -.016 .028 -.ooo -.008 -2.385
.310 -.039 -.621 -.013 .026 .ooo -.008 -.345 Locus:
41001 60001 40599 40750 40582 40575 40354 Total
-1.391 -3.579 -1.016 -.021 -.014 .ooo .ooo -6.021
41001 60001 40599 40750 40582 40575 40354 Total
-.571 -1.210 -2.179 -.005 -.014 -.ooo ,008 -3.971
lS57
-.223 -.725 -.698 -.005 -.013 -.ooo .008 -1.656 Locus:
41001 60001 40599 40750 40582 40575 40354 Total
-1.793 .612 -1.132 .009 -.012 .ooo -.016 -2.332
Model:
-.813 .512 -.812 ,007 -.OlO .ooo -.014 -1.130
Model:
Model:
Model: -.692 -.819 -.472 -.015 -.Oll .ooo .ooo -2.009
MYCLl
Model: -.075 -.450 -.436 -.004 -.Oll -.ooo .008 -.968
lS15
1,
Bale model .2000
.256 .124 -.396 -.OlO .023 ,000 -.008 -.Oll
-.939 -1.471 -.666 -.018 -.013 ,000 .ooo -3.107 Locus:
model
-.585 -1.586 -.762 -.008 -.022 ,000 -.005 -2.968
ALPL
2
-
-543 -1.804 -1.094 ,082 ,006 .ooo -.015 -3.368 PND
AL.
to Loci on Chromosome lp for the Seven Familial Melanoma Pedigrees in Fig. (kla) and Bale (b3a) Models for Recombination Fractions of 0,0.05,0.1,0.2,0.3,
.lOOO DlS47
ET
Model: -.409 .413 -.603 .005 -.008 .ooo -.Oll -.613
.3000
.4000
Family
.oooo
kla -.211 -.708 -.575 ,040 .007 .ooo -.OlO -1.457
-.012 -.197 -.274 .014
-.015 .008 -.098 .003 .OOl ,000 -.OOl -.102
41001 60001 40599 40750 40582 40575 40354 Total
-9.302 -26.926 -8.875 -1.290 -2.809 -16.216 -14.017 -79.435
-.104 -.327 -.076 -.002 -.OOl .ooo .ooo -.516
-.024
41001
.Oll -.ooo -.002 .ooo .ooo -.113
40599 40750 40582 40575 40354 Total
-6.441 -14.150 -11.220 .296 -7.331 -.005 -11.210 -50.061
.075 .166 -.159 -.003 .007 .ooo -.003 .083
.020 .085 -.095 -.OOl .002 .ooo -.OOl .OlO
41001 60001 40599 40750 40582 40575 40354 Total
.381 -5.658 -6.028 -.196 -.229 -1.438 -.302 -13.470
-.176 -.032 -A39 -.004 -.004 .ooo .ooo -.355
-.046 .027 -.064 -.OOl -.OOl .ooo .ooo -.085
41001 60001 40599 40750 40582 40575 40354 Total
-11.386 -9.943 -3.565 -1.403 -.404 -.007 -.OOl -26.709
.039 -.046 -.081 -.OOl -.004 .ooo .003 -.090
,015 -.004 -.020 -.ooo -.OOl .ooo .OOl -.009
40575 40354 Total
-1.856 -5.524 -2.448 -.910 -.322 -4.720 -.813 -22.593
.026 .094 -.166 .OOl -.002 .ooo -.002 -.049
.021 ,025 -.063 .ooo -.ooo .ooo -.ooo -.017
41001 60001 40599 40750 40582 40575 40354 Total
-5.718 .399 -11.245 -1.701 -1.407 -4.915 -13.843 -38.430
41001 60001 40599 40750
Model:
.324 -.669 -.787 -.153 ,051 -.707 -.302 -2.243
Model: .268 -.192 -.398 -.lll .148 -.488 -.299 -1.078
IS57
-3.997 -3.681 -1.136 -1.026 -.362 -.007 -.OOl -10.215
Model:
-2.026 -1.717 -.605 -.477 -.316 -1.022 -.127 -6.289
Model: -1.130 -.930 -.364 -.294 -.294 -.707 .043 -3.676
IS15
Model: -1.640 .319 -2.362 -.561 -.414 - 1.058 -1.243 -6.959
.3000
.4000
-.232 -.208 -.127 -.007 ,093 -.306 -.237 -1.024
-.052 .043 -.013 -.005 .031 -.070 -.054 -.120
-.279 -.581 -.029 .058 -.222 -.002 .018 -1.037
-.065 -.200 .076 ,015 -.052 -.ooo .013 -.213
.078 .192 -.139 -.027 .113 -.149 -.148 -.080
.021 .116 -.126 -.007 .035 -.044 -.042 -.047
b3a -.622 -.896 -.516 -.041 .121 --.810 -.637 -3.401 b3a -.718 -1.255 -.388 ,123 -.567 -.003 -.066 -2.874 b3a .164 .146 -.163 -.063 .174 -.281 -.256 -.279 b3a -1.168 -.998 -.339 -.317 -.201 -.004 -.OOl -3.028
-2.553 -2.286 -.681 -.699 -.315 -.006 -.OOl -6.541
MYCLl
-2.784 .369 -3.529 -.868 -.654 -1.768 -2.188 -11.422
Model: -1.615 -2.643 -1.265 .205 -1.287 -.005 -.400 -7.010
ALPL
Locus:
kla -.071 .229 -.331 .002 -.004 ,000 -.006 -.181
PND
-2.525 -4.187 -2.265 .249 -2.061 -.005 -.848 -11.642
Locus:
.2000
-1.499 -2.512 -1.556 -.221 .018 -1.919 -1.536 -9.225
Locus:
kla .033 -.165 -.198 -.002 -.008 .ooo .006 -.334
-2.486 -4.367 -2.814 -.472 -.I89 -3.147 -2.566 -16.041
Locus:
kla -.382 -.251 -.256 -.009 -.007 .ooo .ooo -.904
DlS47
Locus:
kla ,157 .204 -.229 -.006 .015 .ooo -.006 .135
.lOOO
Locus:
kla -.269 -.754 -.296 -.005 -.015 .ooo -.OOl -1.340
.0500
-.471 -.402 -.216 -.123 -.095 -.002 -.ooo -1.309
-.I12 -.117 -.130 -.028 -.024 -.OOl ,000 -.412
b3a -.386 -.292 -.163 -.118 -.204 -.351 .128 -1.386
-.109 -.061 -.069 -.040 -.099 -.144 ,094 -.428
-.017 ,006 -.018 -.008 -.026 -.034 .031 -.066
-.185 .080 -.580 -.095 -.072 -.157 -.128 -1.137
-.029 ,019 -.215 -.021 -.Oll -.035 -.021 -.319
b3a -.610 ,193 -1.194 -.248 -.184 -.430 -.439 -2.912
EXCLUSION
-60
--
-70
--
-80
1 0
OF
FAMILIAL
MELANOMA
LOCUS
FROM
23
lp.
b
a
20 Genetic
40 Distance
I 60
) 80 (CM) from
: ; 100 120 Chr 1 p telomere
FIG. 2. Exclusion map of a familial melanoma the Kefford model. A horizontal line at a lod value from linkage to MLM.
: 140
160
0
20 Genetic
40 60 80 Distance (CM) from
100 120 Chr 1 p telomere
140
160
locus from regions of lp compiled from the data in Table 2 using (a) the Bale model and (b) of -2 allows visualization of the distances around each of the markers that can be excluded
seven allele systems shown in Table 1. Families were then analyzed using the models of Kefford et al. (1991) and Bale et al. (1989). Table 2 shows the resulting lod scores for various recombination fractions, with the totaled values for each locus plotted against genetic distances (Dracopoli et al., 1991) represented in Fig. 2. Using Bale’s model we can exclude the entire study region from linkage to familial melanoma. Specifically, the combined lod scores for our families are less than -2 for a contiguous region of more than 135 CM (sex average) spanning a region from around 20 CM telomeric to DlS47 to 20 CM proximal to DlS15 (Fig. 2a). For comparison (Fig. 2b), the more stringent nature of Kefford’s model, and some reasonably uninformative lod scores for most families at ALPL, has left uncertainty within this region of exclusion of approximately 28 CM, at the center of which ALPL reaches a lod of -2.385 at a recombination fraction of 0. Another small gap in the exclusion map of less than 12 CM exists between MYCLl and DlS15 using Kefford’s model. In light of the reported linkage to the DlS47/PND region (Bale et al., 1989) particular attention should be paid to the fact that a section of approximately 40 CM around these two loci is excluded from linkage to MLM in our pedigrees using the model of Kefford et al. (1991). The varying extent of these regions of exclusion (Fig. 2) highlights the striking difference between the two models. While all families appear capable of reasonable scores with the Bale model, the Kefford model (with allowances for sporadic cases and age-related, incomplete penetrance) splits the families into two groups; those that do attain significant lod scores and those that always have values close to zero (Table 2). The three largest families (60001, 40599, 41001) consistently yield significant lod scores while the rest give values around zero.
DISCUSSION
A group of six Northern American DNS/MLM families has been reported to have a heritable factor segregating with the PND/DlS47 region of chromosome lp (Bale et al., 1989). The same region has been excluded from linkage in six Dutch (Gruis et al., 1990), three Utah (Cannon-Albright et al., 1990), and eight New South Wales (Kefford et aZ., 1991) pedigrees. All of these kindreds are DNS families, except for those from New South Wales. Kefford et al. (1991) found three types of melanoma families, those with melanoma and dysplastic nevi (not DNS) segregating together, those with MLM and DN, but not segregating together, and one family had four members with melanoma, but no DN. According to their definition, they considered that no individual in their families had DNS as characterized by the presence of greater than 50 atypical moles distributed over sun-exposed and nonexposed sites, where at least one DN had been histologically verified. We present here seven familial melanoma kindreds, only four of which have members with DNS or isolated DN, each of whom also had melanoma. Using the model of Bale et al. (1989) these families exclude -135 CM, approximately half of the short arm of chromosome 1, including the PND/DlS47 and MYCLl regions from linkage to MLM. Using the Kefford model (Kefford et aZ., 1991) an MLM locus is excluded from 15 CM on either side of the PND/ DlS47 region, while 26 CM are excluded distal and 16 CM proximal to MYCLl. A gap of approximately 28 CM, lying either side of ALPL, exists between these two regions although the summed lod scores are still negative. Thus, with the Kefford model we have excluded a smaller region and some regions of uncertainty have opened up within the exclusion map. As demonstrated in the paper of Kefford et al. (1991) and
24
NANCARROW
confirmed herein, this occurs because of a general drop in the magnitude of all lod values. It would appear from our results that non-DNS-associated familial melanoma is not linked to the telomeric portion of chromosome lp and therefore forms a subgroup of melanoma genetically distinct to that reported by Bale et al. (1989). It would also appear that those DNS/MLM families that others have found not to be linked to the PND/DlS47 region (van Haeringen et al., 1989; Gruis et aZ., 1990; Cannon-Albright et al., 1990) may form another subset, and an additional group could potentially arise from multicancer pedigrees, with melanoma being part of the tumor spectrum (Lynch et al., 1981). It appears then that familial melanoma is a heterogeneous disease, with the potential to distinguish three or four subtypes. Indeed, Bale and co-workers indicate that due to possible clinical and genetic heterogeneity, the putative locus to which they found linkage may not be the only MLM-related locus, and it would appear that their locus applies only to a subset of DNS/MLM families (Bale et al., 1990). Clearly linkage studies with non-DNS pedigrees should not be analyzed with the Bale model, or any other DNS-incorporating model. Moreover, in light of the above indications of heterogeneity and the problems of sporadic DN and familial DNS diagnosis, any supposed DNS/MLM pedigrees should also be analyzed with a melanoma-only model. Because of the allowances for incomplete penetrance, age of onset and sporadic melanoma, the Kefford model is well suited to this task. The differences between Figs. 2a and 2b show how great an effect these factors have upon linkage data. It should be noted, however, that the Kefford model, with its high sporadic rate, may only be appropriate for Australian families, in which the lifetime risk of melanoma may be 10 or more times greater than other populations (Balch et al., 1985). Moreover, all models so far considered for MLM are overly simple for such a complex disease phenotype. No model put forward to date has included allowances for residual genetic variation or simultaneous evaluation of epidemiologically important covariates such as sun exposure or skin type. These factors obviously affect the penetrance of the trait differently for each individual, and thus present limitations to linkage analysis. The data presented in this paper show that seven Australian non-DNS, familial melanoma pedigrees, including the three most dense melanoma families so far reported, are not linked to the PND/DlS47 or MYCLl regions of chromosome lp using a realistic model for melanoma. We also illustrate the importance of using a “melanoma only” model to present all MLM linkage data to allow comparisons to be done between data sets. Most importantly, the literature surveyed indicates that familial melanoma may be
ET
AL.
both clinically and genetically heterogeneous, with at least four possible subclasses, and that DNS is not an invariant predisposing or associated factor. ACKNOWLEDGMENTS We thank all the family members, general practitioners, surgeons, and pathologists who so readily cooperated in this study. In particular, we are indebted to Dr. Athel Hockey for many years of helpful collaboration in obtaining samples and demographic information from members of family 60001. The assistance of Fran Sanders and Margaret McJannett in collection of some patient data and blood samples is greatly appreciated. We are grateful to Dr. Robert MacLennan for his help and advice in establishing procedures for patient contact and sample collection. We especially thank Dr. Kay Ellem for encouragement, advice, and critical discussion throughout the course of this project. This work was financially supported by the National Health and Medical Research Council of Australia and the Queensland Cancer Fund. N.H. is a recipient of an R. Douglas Wright award from the NH&MRC.
REFERENCES 1.
BALCH, C. M., SOONG, “Cutaneous Melanoma: wide” (Balch and Milton, delphia.
2.
BALE, S. J., DFLACOPOLI, N. C., GREENE, D. S., AND HOUSMAN, D. E. (1987). Linkage noma and dysplastic nevus syndrome with on human chromosome lp. Cytogenet. Cell.
3.
BALE, S. J., DRACOPOLI, N. C., TUCKER, M. A., CLARK, W. H., FRASER, M. C., STANGER, B. Z., GREEN, P., DONIS-KELLER, H., HOUSMAN, D. E., AND GREENE, M. H. (1989). Mapping the gene for hereditary cutaneous malignant melanoma-dysplastic nevus to chromosome lp. N. Engl. J. Med. 320: 13671372.
4.
BALE, S. J., TUCKER, M. A., AND DRACOPOLI, the Editor: N. Engl. J. Med. 322: 854.
5.
CANNON-ALBRIGHT, L. A., GOLDGAR, D. E., WRIGHT, E. C., R~RCO, A., JOST, M., MEYER, L. J., PIEPKORN, M., ZONE, J. J., AND SKOLNICK, M. H. (1990). Evidence against the reported linkage of the cutaneous melanoma-dysplastic nevus syndrome locus to chromosome 1~36. Am. J. Hum. Genet. 46: 912-918. CHURCH, G. M., AND GILBERT, W. (1984). Genomic sequencing. Proc. Natl. Acad. Sci. USA 81: 1991-1995.
6.
S. J., AND SHAW, H. M. (1985). In Characteristics and Results WorldEds.), pp. 507-20, Lippincott, PhilaM. H., GERHARD, analysis of melapolymorphic loci Genet. 46: 575.
N. C. (1990).
To
7.
CLARK, W. H., WELCH, H. M., AND GREENE, M. H. (1978). Origin of familial malignant melanomas from heritable melanocytic lesions. Arch. Dermatol. 114: 732-738.
8.
Consensus Development Conference (1984). Precursors to malignant melanoma. J. Am. Med. Assoc. 251: 1864-1866. DONIS-KELLER, H., GREEN, P., HELMS, C., CARTINHOUR, S., WEIFFENBACH, B., STEPHENS, K., KEITH, T. P., BOWDEN, D. W., SMITH, D. R., LANDER, E. S., BOTSTEIN, D., AKOTS, G., REDIKER, K. S., GRA~IUS, T., BROWN, V. A., RISING, M. B., PARKER, C., POWERS, J. A., AND WA?T, D. E. (1987). A genetic linkage map of the human genome. Cell 51: 319-337.
9.
10.
DRACOPOLI, N. C., O’CONNELL, P., ELSNER, T. I., LALOUEL, J., WHITE, R. L., BUETOW, K. H., NISHIMURA, D. Y., MURRAY, J. C., HELMS, C., MISHRA, S. K., DONIS-KELLER, H., HALL, J. M., LEE, M. K., KING, M., A?-pwoo~, J., MORTON, N. E., ROBSON, E. B., MAHTANI, M., WILLARD, H. W., ROYLE, N. J., PATEL, I., JEFFREYS, A. J., VERGA, JENKINS, T., WEBER, J. L., MITCHELL, A. L., AND BALE, A. E. (1991). The
EXCLUSION
11.
12.
13.
14.
15.
16.
17.
OF
FAMILIAL
CEPH consortium linkage map of human chromosome 1. Genomics 9: 686-700. DRACOPOLI, N. C., STANGER, B. Z., ITO, C. Y., CALL, K. M., LINCOLN, S. E., LANDER, E. S., AND HOUSMAN, D. E. (1988). A genetic linkage map of 27 loci from PND to FY on the short arm of human chromosome 1. Am. J. Hum. Genet. 43: 462470. FEINBERG, A. P., AND VOGELSTEIN, B. (1983). A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem. 132: 6-13. FEINBERG, A. P., AND VOGELSTEIN, B. (1984). Addendum: A technique for radiolabeling DNA Restriction endonuclease fragments to high specific activity. An&. Biochem. 137: 266267. GREENE, M. H., GOLDIN, L. R., CLARK, W. H., LOVFUEN, E., KRAEMER, K. H., TUCKER, M. A., ELDER, D. E., FRASER, M. C., AND ROWE, S. (1983). Familial cutaneous malignant melanoma: Autosomal dominant trait possibly linked to the RH locus. Proc. Natl. Acad. Sci. USA 80: 6071-6075. GRUIS, N. A., BERGMAN, W., AND FRANTS, R. R. (1990). Locus for susceptibility to melanoma on chromosome lp. N. Engl. J. Med. 322: 853-854. KEFFORD, R. F., SALMON, J., SHAW, H. M., DONALD, J. A., AND MCCARTHY, W. H. (1991). Hereditary melanoma in Australia: variable association with dysplastic nevi and absence of genetic linkage to chromosome lp. Cancer Genet. Cytogenet. 51: 45-55. LYNCH, H. T., FRICHOT, B. C., AND LYNCH, J. F. (1978). Familial atypical multiple mole melanoma syndrome. J. Med. Genet. 15: 352-356.
MELANOMA 18.
19.
20.
21.
22.
LOCUS
FROM
lp.
25
LYNCH, H. T., FLJSARO, R. M., PESTER, J., OOSTERHUIS, J. A., WENT, L. N., RUMKE, P., NEERING, H., AND LYNCH, J. F. (1981). Tumour spectrum in the FAMMM syndrome. Br. J. Cancer 44: 553-560. MILLER, S. A., DYKES, D. D., AND POLESKY, H. F. (1988). A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res. 16: 1215. NAKAMURA, Y., LEPPERT, M., O’CONNELL, P., WOLFF, R., HOLM, T., CULVER, M., MARTIN, C., FUJIMOTO, E., HOFF, M., KUMLIN, E., AND WHITE, R. (1987). Variable number tandem repeat (VNTR) markers for human gene mapping. Science 235: 1616-1622. NEMER, M., SIROIS, D., AND DROUIN, J. (1986). XhoI polymorphism at the human pronatriodilatin (hPND) gene locus. Nucleic Acids Res. 14: 8696. Orr, J. (1976). A computer program for linkage analysis of general pedigree data. Am. J. Hum. Genet. 28: 528-529.
23.
RAMSAY, R. G., CHEN, P., IMMRAY, F. P., KIDSON, C., LAVIN, M. F., AND HOCKEY, A. (1982). Familial melanoma associated with dominant ultraviolet radiation sensitivity. Cancer Res. 42: 2909-2912.
24.
VAN HAERINGEN, A., BERGMAN, W., NELEN, M. R., VAN DER KOOIJ-MEIJS, E., HENDRIKSE, I., WLJNEN, J. T., KHAN, P. M., KLASEN, E. C., AND FFUNTS, R. R. (1989). Exclusion of the dysplastic naevus syndrome (DNS) locus from the short arm of chromosome 1 by linkage studies in Dutch families. Genomics 5: 61-64.
25.
WEISS, M. J., SPIELMAN, R. S., AND HARRIS, H. (1987). high-frequency RFLP at the human liver/bone/kidney-type alkaline phosphatase locus. Nucleic Acids Res. 15: 860.
A
12,
18-25
(19%)
Exclusion of the Familial Melanoma Locus (MLM) from the PND/ DlS47 and MYCLI Regions of Chromosome Arm Ip in 7 Australian Pedigrees DEREK J. NANCARROW, JANE M. PALMER, MARILYN K. WALTERS, BEVERLEY M. KERR, GREGORY J. HAFNER, LUKE GARSKE, G. RODERICK MCLEOD,* AND NICHOLAS K. HAYWARD’ Queensland Cancer Fund Research Unit, Joint Experimental Oncology Program, Queensland Institute of Medical Research, Herston, 4029, Australia; and *Queensland Melanoma Project, Princess Alexandra Hospital, Wooloongabba, 4 702, Australia ReceivedJune
17. 1991;revisedAugust23,1991
velopment Conference, 1984), in which multiple, atypical moles are frequently found in association with malignant melanoma. Thus DNS formed the physical marker for familial melanoma carriers based on this circumstantial evidence. Using the combination of melanoma and DNS as the disease phenotype, Greene et al. (1983) found a suggestion of loose linkage (Z,, = 2 at 0 = 0.3) in 14 dysplastic nevus syndrome/familial melanoma (DNS/MLM) families to the rhesus blood group locus on the short arm of chromosome one. Four years later Bale et al. (1987) verified this linkage to rhesus and several other markers on lp. In 1989 the same group claimed to have localized this putative DNS/MLM gene, designated CMM, to a region between the anonymous DNA marker DlS47 and the gene for pronatriodilatin (PND), 36 CM (Dracopoli et aZ., 1991) telomerit to rhesus, in six families (Bale et al., 1989). Coinciding with this report, van Haeringen et al. (1989) published data that claimed to exclude all of the short arm of chromosome 1 in six Dutch DNS or DNS/MLM families, and in 1990, the same group followed up these findings (in the same six families) with a more specific exclusion, using the DlS47 and PND probes (Gruis et al., 1990). Two other groups have published data excluding the DlS47/PND region from linkage to CMM (DNS/ MLM) using the same model for DNS/MLM as Bale et al. (1989). Cannon- Albright et al. (1990) presented exclusion data for this region in three Utah DNS/ MLM pedigrees, while Kefford et al. (1991) excluded this region in eight Australian kindreds. Both of these groups openly questioned combining MLM and DNS to study linkage and each paper included analyses of MLM alone. Cannon-Albright et al. (1990) discussed the subjective definition of DNS described by Greene et al. (1983) and the subjective nature of the diagnosis. Several of the Australian families presented by Kefford et al. (1991) showed little or no history of dys-
Familial melanoma (MLM) is sometimes found associated with the dysplastic nevus syndrome (DNS). Considerable controversy exists over the possible assignment of a cutane-
ous malignant melanomafdysplastic nevus gene, designated CMM, to the distal short arm of chromosome 1, linked to the PND and DlS47 loci. To date, no support for linkage of MLM alone to these markers has been found; likewise no study has been able to exclude the entire region between PND and DlS47 from linkage to MLM. We have carried out linkage studies between markers on lp and MLM in seven Australian kindreds; three of these are the largest reported worldwide. We have been able to exclude localization of an MLM gene from a 40-CM region that spans the interval between DlS47 and PND and extends approximately 15 CM on either side of these markers. In addition, we can exclude a region of about 20 CM around the MYCLl/DlSB? loci. Q 1992 Academic press, IX.
INTRODUCTION
When analyzing a genetic disease it is useful to have a physical marker that appears to be associated with the disease in question, a precursor, or an early symptom. This is particularly important for identifying family members who are carriers in familial diseasesthat are codominant, have late or variable age of onset, or have a reduced penetrance. Familial melanoma, although apparently dominant, has both an age-relatedonset and a reducedpenetrance, asdemonstrated by skipped generations in pedigrees worldwide. Lynch et al. (1978) and Clark et al. (1978) independently characterized what was later to be called the dysplastic nevus syndrome (DNS) (Consensus De1 To whom correspondence should be addressed Institute of Medical Research, Herston Rd., Queensland, Australia.
ossa-7543/92 $3.00 Copyright 0 1992 by Academic Press, All rights of reproduction in any form
at Queensland Herston, 4029,
18 Inc.
reserved.
EXCLUSION
U-IO-24
25-5-19 167-15 I
OF
FAMILIAL
MELANOMA
361-22 21-l-15 I
LOCUS
n-9-u
w3
FROM
15-Z-39 30-3-70 I
M-I-30 u-11-80 I
026
V.8
19
lp.
031
an
036
wn
FIG. 1. Pedigrees of the seven familial melanoma kindreds used in this study. (m) Affected male; (0) unaffected female; (EJ) deceased unaffected male; (h) deceased affected female. Only the numbered individuals were available for genotyping in this study. Alleles for the DlS47 locus and BglI/XhoI haplotypes of the PND locus respectively are given below each family member. Dates, in descending position above and left of each person, are date of birth, date of first melanoma, and date of death.
plastic nevus syndrome or surgical removal of histologically categorized dysplastic nevi. This led them to introduce a linkage model for melanoma alone, based on data from the Sydney Melanoma Unit. This model takes into consideration age of onset, sporadic mela-
noma, and incomplete penetrance (factors all seen in familial melanoma kindreds but usually left out of previous models). We present here a comprehensive analysis of the p32-p36 region of chromosome 1, carried out on six
20
NANCARROW
8/D a/a 014
B/D r/b 004
YY */a 003
B/6 r/o 042
ET
AL.
WE alo 043
m9
oaYm5mD FIG.
DO1
l-Continued
familial melanoma pedigrees collected in Queensland and one in Western Australia, three of which contain more living affecteds than any other kindred thus far presented. We have chosen to present our data using the model of Kefford et al. (1991) for melanoma alone (O-15 years 1% penetrance, 0% sporadic; 16-25 years 30% penetrance, 1% sporadic; 26-45 years 50% penetrance, 3% sporadic; 46 years and over 80% penetrance, 5% sporadic; with a gene frequency of 0.003) for several reasons; our close geographical proximity to families collected by the Kefford group, because we too have non-DNS families and because the data used to calculate parameters of the model were estimated
from a population based survey. We have, however, included an analysis with the model of Bale et al. (1989) (99% penetrance for all people over 20 years with a gene frequency of 0.001 and no allowance for sporadic melanoma) strictly for comparison with existing reports. The data presented here are the first to exclude the entire region between the PND and DlS47 loci from linkage to MLM alone, with both these models. METHODS DNA from 40 patients with histologically confirmed melanoma and 57 unaffected members of 7 fa-
EXCLUSION
TABLE
OF
FAMILIAL
1
Allele Sizes and Frequencies for Probes Used to Study Possible Linkage of Familial Melanoma to Loci on lp Probe CRI-336 pJAll0
LOCUS
DlS47 PND
Enzyme RSUI
xhd BglI
pLBK8B/E5
ALPL
BClI
pYNZ2
DlS57
RSd
PLmYc
MYCLl
EcoRI
pl-11B
DlS15
BglII
Allele size (kb)
Allele frequency
VNTR 12 10 10 6 7.4 4.3, 3.1 2.5
0.1666” 0.46 0.54 0.87 0.13 0.67 0.33 0.3334 0.3333 0.3333 0.44 0.56 0.56 0.30 0.14
1.8
1.5 10 6.6 19 17 14.8
Reference (9)
(21) (11)
6’5) POP (11) (11)
“To enable appropriate analysis with LIPED we have nominated six alleles of equal frequency. b Since we only observed three of the five alleles seen by Nakamura et al. (20) we have adjusted the frequencies accordingly.
milial melanoma pedigrees (Fig. 1) was obtained from EBV-transformed lymphocytes using the method of Miller et al. (1988). Only family 60001 has been published previously, in part (family B in Ramsay et al. (1982)). The other six families were obtained through the Queensland Melanoma Project (Princess Alexandra Hospital, Brisbane) and were selected by establishing whether probands had at least two other family members with melanoma, still living. Thus, in our study, MLM is defined as three or more living first- or second-degree relatives with histologically confirmed melanoma. In three families (40575, 40354, and 40750) there were no cases (either melanoma patients or unaffected relatives) of classical dysplastic nevus syndrome (Lynch et aZ., 1978; Clark et al., 1978) or isolated excision of a single (sporadic) nevus histopathologically categorized as dysplastic. Only one member (213) of family 40582 and two members (009, 021) of family 40599 had classical DNS. Each of these patients also had melanoma. A nurse or general practitioner asked all family members whether they had at any time had any pigmented lesions surgically removed. Histology reports were obtained for each of these lesions and the diagnosis noted. Histologically classified dysplastic nevi were excised from patients 212 in family 40582; 007 and 005 (two) in family 41001; and 002, 005, 017 (two), and 021 (four) in family 60001. Once again, all instances of surgical removal of DN were in family members who also had histopathologically confirmed melanoma. In contrast, there were no family members
MELANOMA
LOCUS
FROM
lp.
21
without melanoma who had DNS or had surgical excision of an isolated DN. All nevi excised from nonaffected family members (and some of the other nevi removed from affected cases) were of the benign compound, intradermal, or junctional type. For each RFLP studied, DNA from all 97 individuals was cleaved with the appropriate restriction enzyme, electrophoresed in 0.8-1.2% agarose gels (depending on the expected allele sizes), then capillary blotted onto nylon membranes using 10X SSC. Prehybridization, hybridization, and stringency washes were carried out according to the method of Church and Gilbert (1984). Probe DNA was labeled by the random priming method of Feinberg and Vogelstein (1983, 1984). The probes used in this study were chosen to correspond with those used in previously reported linkage analyses between MLM or CMM (DNS/MLM) and chromosome lp markers, and are listed in Table 1, along with their locus name, the enzyme used to generate the RFLP, and the corresponding allele frequencies. The latter were all derived from Caucasian populations (see Table 1 for references) and should thus be appropriate for our pedigrees, which are all AngloSaxon in origin. In the case of the PND locus two enzyme systems were characterised, XhoI and BgZI, then haplotyped for linkage analysis. The hypervariability of DlS47 and the inability to uniquely characterize alleles across families combined with the limitation of computer analysis to a maximum of six alleles at any one locus, resulted in the nomination of six alleles of equal frequency in the analysis of this locus; four for first generation alleles and the remaining two encompassing all spouse alleles. For DlS57, we only observed three alleles, where previously five had been reported (Nakamura et al., 1987), and gave these alleles equal frequencies. All linkage analyses were performed with the computer program LIPED (Ott, 1976), with TABLOID2 compiling these data directly to form Table 2. Two models were used to represent the mode of inheritance of familial melanoma; that of Kefford et al. (1991) and that of Bale et al. (1989). We chose not to use the only other model for MLM/DNS so far proposed (Gruis et cd., 1990) since it does not incorporate variable age of onset, reduced penetrances, or sporadic incidence of disease. All genetic distances discussed in this paper are those of sex average taken from the latest CEPH Consortium Linkage Map (Table 4 in Dracopoli et al. (1991)). RESULTS
Each of the 97 DNA samples (numbered individuals) obtained from the seven familial melanoma kindreds illustrated in Fig. 1 was characterized for the ’ A set of computer to compress LIPED
programs developed in this laboratory output; available upon request.
used
22
NANCARROW
TABLE Lod Scores for Linkage Analyzed with the Kefford and 0.4 Kefford Family
.oooo
.0500 Locus:
41001 60001 40599 40750 40582 40575 40354 Total
-2.184 -7.197 -2.152 ,140 -.005 ,000 -.017 -12.015
-.913 -2.895 -1.500 .109 .002 .ooo -.016 -5.213 Locus:
41001 60001 40599 40750 40582 40575 40354 Total
-1.275 -4.634 -3.244 -.013 -.027 .ooo -.012 -9.205
-.852 -2.399 -1.223 -.OlO -.025 .ooo -.008 -4.517 Locus:
41001 60001 40599 40750 40582 40575 40354 Total
.366 -.602 -2.153 -.016 .028 -.ooo -.008 -2.385
.310 -.039 -.621 -.013 .026 .ooo -.008 -.345 Locus:
41001 60001 40599 40750 40582 40575 40354 Total
-1.391 -3.579 -1.016 -.021 -.014 .ooo .ooo -6.021
41001 60001 40599 40750 40582 40575 40354 Total
-.571 -1.210 -2.179 -.005 -.014 -.ooo ,008 -3.971
lS57
-.223 -.725 -.698 -.005 -.013 -.ooo .008 -1.656 Locus:
41001 60001 40599 40750 40582 40575 40354 Total
-1.793 .612 -1.132 .009 -.012 .ooo -.016 -2.332
Model:
-.813 .512 -.812 ,007 -.OlO .ooo -.014 -1.130
Model:
Model:
Model: -.692 -.819 -.472 -.015 -.Oll .ooo .ooo -2.009
MYCLl
Model: -.075 -.450 -.436 -.004 -.Oll -.ooo .008 -.968
lS15
1,
Bale model .2000
.256 .124 -.396 -.OlO .023 ,000 -.008 -.Oll
-.939 -1.471 -.666 -.018 -.013 ,000 .ooo -3.107 Locus:
model
-.585 -1.586 -.762 -.008 -.022 ,000 -.005 -2.968
ALPL
2
-
-543 -1.804 -1.094 ,082 ,006 .ooo -.015 -3.368 PND
AL.
to Loci on Chromosome lp for the Seven Familial Melanoma Pedigrees in Fig. (kla) and Bale (b3a) Models for Recombination Fractions of 0,0.05,0.1,0.2,0.3,
.lOOO DlS47
ET
Model: -.409 .413 -.603 .005 -.008 .ooo -.Oll -.613
.3000
.4000
Family
.oooo
kla -.211 -.708 -.575 ,040 .007 .ooo -.OlO -1.457
-.012 -.197 -.274 .014
-.015 .008 -.098 .003 .OOl ,000 -.OOl -.102
41001 60001 40599 40750 40582 40575 40354 Total
-9.302 -26.926 -8.875 -1.290 -2.809 -16.216 -14.017 -79.435
-.104 -.327 -.076 -.002 -.OOl .ooo .ooo -.516
-.024
41001
.Oll -.ooo -.002 .ooo .ooo -.113
40599 40750 40582 40575 40354 Total
-6.441 -14.150 -11.220 .296 -7.331 -.005 -11.210 -50.061
.075 .166 -.159 -.003 .007 .ooo -.003 .083
.020 .085 -.095 -.OOl .002 .ooo -.OOl .OlO
41001 60001 40599 40750 40582 40575 40354 Total
.381 -5.658 -6.028 -.196 -.229 -1.438 -.302 -13.470
-.176 -.032 -A39 -.004 -.004 .ooo .ooo -.355
-.046 .027 -.064 -.OOl -.OOl .ooo .ooo -.085
41001 60001 40599 40750 40582 40575 40354 Total
-11.386 -9.943 -3.565 -1.403 -.404 -.007 -.OOl -26.709
.039 -.046 -.081 -.OOl -.004 .ooo .003 -.090
,015 -.004 -.020 -.ooo -.OOl .ooo .OOl -.009
40575 40354 Total
-1.856 -5.524 -2.448 -.910 -.322 -4.720 -.813 -22.593
.026 .094 -.166 .OOl -.002 .ooo -.002 -.049
.021 ,025 -.063 .ooo -.ooo .ooo -.ooo -.017
41001 60001 40599 40750 40582 40575 40354 Total
-5.718 .399 -11.245 -1.701 -1.407 -4.915 -13.843 -38.430
41001 60001 40599 40750
Model:
.324 -.669 -.787 -.153 ,051 -.707 -.302 -2.243
Model: .268 -.192 -.398 -.lll .148 -.488 -.299 -1.078
IS57
-3.997 -3.681 -1.136 -1.026 -.362 -.007 -.OOl -10.215
Model:
-2.026 -1.717 -.605 -.477 -.316 -1.022 -.127 -6.289
Model: -1.130 -.930 -.364 -.294 -.294 -.707 .043 -3.676
IS15
Model: -1.640 .319 -2.362 -.561 -.414 - 1.058 -1.243 -6.959
.3000
.4000
-.232 -.208 -.127 -.007 ,093 -.306 -.237 -1.024
-.052 .043 -.013 -.005 .031 -.070 -.054 -.120
-.279 -.581 -.029 .058 -.222 -.002 .018 -1.037
-.065 -.200 .076 ,015 -.052 -.ooo .013 -.213
.078 .192 -.139 -.027 .113 -.149 -.148 -.080
.021 .116 -.126 -.007 .035 -.044 -.042 -.047
b3a -.622 -.896 -.516 -.041 .121 --.810 -.637 -3.401 b3a -.718 -1.255 -.388 ,123 -.567 -.003 -.066 -2.874 b3a .164 .146 -.163 -.063 .174 -.281 -.256 -.279 b3a -1.168 -.998 -.339 -.317 -.201 -.004 -.OOl -3.028
-2.553 -2.286 -.681 -.699 -.315 -.006 -.OOl -6.541
MYCLl
-2.784 .369 -3.529 -.868 -.654 -1.768 -2.188 -11.422
Model: -1.615 -2.643 -1.265 .205 -1.287 -.005 -.400 -7.010
ALPL
Locus:
kla -.071 .229 -.331 .002 -.004 ,000 -.006 -.181
PND
-2.525 -4.187 -2.265 .249 -2.061 -.005 -.848 -11.642
Locus:
.2000
-1.499 -2.512 -1.556 -.221 .018 -1.919 -1.536 -9.225
Locus:
kla .033 -.165 -.198 -.002 -.008 .ooo .006 -.334
-2.486 -4.367 -2.814 -.472 -.I89 -3.147 -2.566 -16.041
Locus:
kla -.382 -.251 -.256 -.009 -.007 .ooo .ooo -.904
DlS47
Locus:
kla ,157 .204 -.229 -.006 .015 .ooo -.006 .135
.lOOO
Locus:
kla -.269 -.754 -.296 -.005 -.015 .ooo -.OOl -1.340
.0500
-.471 -.402 -.216 -.123 -.095 -.002 -.ooo -1.309
-.I12 -.117 -.130 -.028 -.024 -.OOl ,000 -.412
b3a -.386 -.292 -.163 -.118 -.204 -.351 .128 -1.386
-.109 -.061 -.069 -.040 -.099 -.144 ,094 -.428
-.017 ,006 -.018 -.008 -.026 -.034 .031 -.066
-.185 .080 -.580 -.095 -.072 -.157 -.128 -1.137
-.029 ,019 -.215 -.021 -.Oll -.035 -.021 -.319
b3a -.610 ,193 -1.194 -.248 -.184 -.430 -.439 -2.912
EXCLUSION
-60
--
-70
--
-80
1 0
OF
FAMILIAL
MELANOMA
LOCUS
FROM
23
lp.
b
a
20 Genetic
40 Distance
I 60
) 80 (CM) from
: ; 100 120 Chr 1 p telomere
FIG. 2. Exclusion map of a familial melanoma the Kefford model. A horizontal line at a lod value from linkage to MLM.
: 140
160
0
20 Genetic
40 60 80 Distance (CM) from
100 120 Chr 1 p telomere
140
160
locus from regions of lp compiled from the data in Table 2 using (a) the Bale model and (b) of -2 allows visualization of the distances around each of the markers that can be excluded
seven allele systems shown in Table 1. Families were then analyzed using the models of Kefford et al. (1991) and Bale et al. (1989). Table 2 shows the resulting lod scores for various recombination fractions, with the totaled values for each locus plotted against genetic distances (Dracopoli et al., 1991) represented in Fig. 2. Using Bale’s model we can exclude the entire study region from linkage to familial melanoma. Specifically, the combined lod scores for our families are less than -2 for a contiguous region of more than 135 CM (sex average) spanning a region from around 20 CM telomeric to DlS47 to 20 CM proximal to DlS15 (Fig. 2a). For comparison (Fig. 2b), the more stringent nature of Kefford’s model, and some reasonably uninformative lod scores for most families at ALPL, has left uncertainty within this region of exclusion of approximately 28 CM, at the center of which ALPL reaches a lod of -2.385 at a recombination fraction of 0. Another small gap in the exclusion map of less than 12 CM exists between MYCLl and DlS15 using Kefford’s model. In light of the reported linkage to the DlS47/PND region (Bale et al., 1989) particular attention should be paid to the fact that a section of approximately 40 CM around these two loci is excluded from linkage to MLM in our pedigrees using the model of Kefford et al. (1991). The varying extent of these regions of exclusion (Fig. 2) highlights the striking difference between the two models. While all families appear capable of reasonable scores with the Bale model, the Kefford model (with allowances for sporadic cases and age-related, incomplete penetrance) splits the families into two groups; those that do attain significant lod scores and those that always have values close to zero (Table 2). The three largest families (60001, 40599, 41001) consistently yield significant lod scores while the rest give values around zero.
DISCUSSION
A group of six Northern American DNS/MLM families has been reported to have a heritable factor segregating with the PND/DlS47 region of chromosome lp (Bale et al., 1989). The same region has been excluded from linkage in six Dutch (Gruis et al., 1990), three Utah (Cannon-Albright et al., 1990), and eight New South Wales (Kefford et aZ., 1991) pedigrees. All of these kindreds are DNS families, except for those from New South Wales. Kefford et al. (1991) found three types of melanoma families, those with melanoma and dysplastic nevi (not DNS) segregating together, those with MLM and DN, but not segregating together, and one family had four members with melanoma, but no DN. According to their definition, they considered that no individual in their families had DNS as characterized by the presence of greater than 50 atypical moles distributed over sun-exposed and nonexposed sites, where at least one DN had been histologically verified. We present here seven familial melanoma kindreds, only four of which have members with DNS or isolated DN, each of whom also had melanoma. Using the model of Bale et al. (1989) these families exclude -135 CM, approximately half of the short arm of chromosome 1, including the PND/DlS47 and MYCLl regions from linkage to MLM. Using the Kefford model (Kefford et aZ., 1991) an MLM locus is excluded from 15 CM on either side of the PND/ DlS47 region, while 26 CM are excluded distal and 16 CM proximal to MYCLl. A gap of approximately 28 CM, lying either side of ALPL, exists between these two regions although the summed lod scores are still negative. Thus, with the Kefford model we have excluded a smaller region and some regions of uncertainty have opened up within the exclusion map. As demonstrated in the paper of Kefford et al. (1991) and
24
NANCARROW
confirmed herein, this occurs because of a general drop in the magnitude of all lod values. It would appear from our results that non-DNS-associated familial melanoma is not linked to the telomeric portion of chromosome lp and therefore forms a subgroup of melanoma genetically distinct to that reported by Bale et al. (1989). It would also appear that those DNS/MLM families that others have found not to be linked to the PND/DlS47 region (van Haeringen et al., 1989; Gruis et aZ., 1990; Cannon-Albright et al., 1990) may form another subset, and an additional group could potentially arise from multicancer pedigrees, with melanoma being part of the tumor spectrum (Lynch et al., 1981). It appears then that familial melanoma is a heterogeneous disease, with the potential to distinguish three or four subtypes. Indeed, Bale and co-workers indicate that due to possible clinical and genetic heterogeneity, the putative locus to which they found linkage may not be the only MLM-related locus, and it would appear that their locus applies only to a subset of DNS/MLM families (Bale et al., 1990). Clearly linkage studies with non-DNS pedigrees should not be analyzed with the Bale model, or any other DNS-incorporating model. Moreover, in light of the above indications of heterogeneity and the problems of sporadic DN and familial DNS diagnosis, any supposed DNS/MLM pedigrees should also be analyzed with a melanoma-only model. Because of the allowances for incomplete penetrance, age of onset and sporadic melanoma, the Kefford model is well suited to this task. The differences between Figs. 2a and 2b show how great an effect these factors have upon linkage data. It should be noted, however, that the Kefford model, with its high sporadic rate, may only be appropriate for Australian families, in which the lifetime risk of melanoma may be 10 or more times greater than other populations (Balch et al., 1985). Moreover, all models so far considered for MLM are overly simple for such a complex disease phenotype. No model put forward to date has included allowances for residual genetic variation or simultaneous evaluation of epidemiologically important covariates such as sun exposure or skin type. These factors obviously affect the penetrance of the trait differently for each individual, and thus present limitations to linkage analysis. The data presented in this paper show that seven Australian non-DNS, familial melanoma pedigrees, including the three most dense melanoma families so far reported, are not linked to the PND/DlS47 or MYCLl regions of chromosome lp using a realistic model for melanoma. We also illustrate the importance of using a “melanoma only” model to present all MLM linkage data to allow comparisons to be done between data sets. Most importantly, the literature surveyed indicates that familial melanoma may be
ET
AL.
both clinically and genetically heterogeneous, with at least four possible subclasses, and that DNS is not an invariant predisposing or associated factor. ACKNOWLEDGMENTS We thank all the family members, general practitioners, surgeons, and pathologists who so readily cooperated in this study. In particular, we are indebted to Dr. Athel Hockey for many years of helpful collaboration in obtaining samples and demographic information from members of family 60001. The assistance of Fran Sanders and Margaret McJannett in collection of some patient data and blood samples is greatly appreciated. We are grateful to Dr. Robert MacLennan for his help and advice in establishing procedures for patient contact and sample collection. We especially thank Dr. Kay Ellem for encouragement, advice, and critical discussion throughout the course of this project. This work was financially supported by the National Health and Medical Research Council of Australia and the Queensland Cancer Fund. N.H. is a recipient of an R. Douglas Wright award from the NH&MRC.
REFERENCES 1.
BALCH, C. M., SOONG, “Cutaneous Melanoma: wide” (Balch and Milton, delphia.
2.
BALE, S. J., DFLACOPOLI, N. C., GREENE, D. S., AND HOUSMAN, D. E. (1987). Linkage noma and dysplastic nevus syndrome with on human chromosome lp. Cytogenet. Cell.
3.
BALE, S. J., DRACOPOLI, N. C., TUCKER, M. A., CLARK, W. H., FRASER, M. C., STANGER, B. Z., GREEN, P., DONIS-KELLER, H., HOUSMAN, D. E., AND GREENE, M. H. (1989). Mapping the gene for hereditary cutaneous malignant melanoma-dysplastic nevus to chromosome lp. N. Engl. J. Med. 320: 13671372.
4.
BALE, S. J., TUCKER, M. A., AND DRACOPOLI, the Editor: N. Engl. J. Med. 322: 854.
5.
CANNON-ALBRIGHT, L. A., GOLDGAR, D. E., WRIGHT, E. C., R~RCO, A., JOST, M., MEYER, L. J., PIEPKORN, M., ZONE, J. J., AND SKOLNICK, M. H. (1990). Evidence against the reported linkage of the cutaneous melanoma-dysplastic nevus syndrome locus to chromosome 1~36. Am. J. Hum. Genet. 46: 912-918. CHURCH, G. M., AND GILBERT, W. (1984). Genomic sequencing. Proc. Natl. Acad. Sci. USA 81: 1991-1995.
6.
S. J., AND SHAW, H. M. (1985). In Characteristics and Results WorldEds.), pp. 507-20, Lippincott, PhilaM. H., GERHARD, analysis of melapolymorphic loci Genet. 46: 575.
N. C. (1990).
To
7.
CLARK, W. H., WELCH, H. M., AND GREENE, M. H. (1978). Origin of familial malignant melanomas from heritable melanocytic lesions. Arch. Dermatol. 114: 732-738.
8.
Consensus Development Conference (1984). Precursors to malignant melanoma. J. Am. Med. Assoc. 251: 1864-1866. DONIS-KELLER, H., GREEN, P., HELMS, C., CARTINHOUR, S., WEIFFENBACH, B., STEPHENS, K., KEITH, T. P., BOWDEN, D. W., SMITH, D. R., LANDER, E. S., BOTSTEIN, D., AKOTS, G., REDIKER, K. S., GRA~IUS, T., BROWN, V. A., RISING, M. B., PARKER, C., POWERS, J. A., AND WA?T, D. E. (1987). A genetic linkage map of the human genome. Cell 51: 319-337.
9.
10.
DRACOPOLI, N. C., O’CONNELL, P., ELSNER, T. I., LALOUEL, J., WHITE, R. L., BUETOW, K. H., NISHIMURA, D. Y., MURRAY, J. C., HELMS, C., MISHRA, S. K., DONIS-KELLER, H., HALL, J. M., LEE, M. K., KING, M., A?-pwoo~, J., MORTON, N. E., ROBSON, E. B., MAHTANI, M., WILLARD, H. W., ROYLE, N. J., PATEL, I., JEFFREYS, A. J., VERGA, JENKINS, T., WEBER, J. L., MITCHELL, A. L., AND BALE, A. E. (1991). The
EXCLUSION
11.
12.
13.
14.
15.
16.
17.
OF
FAMILIAL
CEPH consortium linkage map of human chromosome 1. Genomics 9: 686-700. DRACOPOLI, N. C., STANGER, B. Z., ITO, C. Y., CALL, K. M., LINCOLN, S. E., LANDER, E. S., AND HOUSMAN, D. E. (1988). A genetic linkage map of 27 loci from PND to FY on the short arm of human chromosome 1. Am. J. Hum. Genet. 43: 462470. FEINBERG, A. P., AND VOGELSTEIN, B. (1983). A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem. 132: 6-13. FEINBERG, A. P., AND VOGELSTEIN, B. (1984). Addendum: A technique for radiolabeling DNA Restriction endonuclease fragments to high specific activity. An&. Biochem. 137: 266267. GREENE, M. H., GOLDIN, L. R., CLARK, W. H., LOVFUEN, E., KRAEMER, K. H., TUCKER, M. A., ELDER, D. E., FRASER, M. C., AND ROWE, S. (1983). Familial cutaneous malignant melanoma: Autosomal dominant trait possibly linked to the RH locus. Proc. Natl. Acad. Sci. USA 80: 6071-6075. GRUIS, N. A., BERGMAN, W., AND FRANTS, R. R. (1990). Locus for susceptibility to melanoma on chromosome lp. N. Engl. J. Med. 322: 853-854. KEFFORD, R. F., SALMON, J., SHAW, H. M., DONALD, J. A., AND MCCARTHY, W. H. (1991). Hereditary melanoma in Australia: variable association with dysplastic nevi and absence of genetic linkage to chromosome lp. Cancer Genet. Cytogenet. 51: 45-55. LYNCH, H. T., FRICHOT, B. C., AND LYNCH, J. F. (1978). Familial atypical multiple mole melanoma syndrome. J. Med. Genet. 15: 352-356.
MELANOMA 18.
19.
20.
21.
22.
LOCUS
FROM
lp.
25
LYNCH, H. T., FLJSARO, R. M., PESTER, J., OOSTERHUIS, J. A., WENT, L. N., RUMKE, P., NEERING, H., AND LYNCH, J. F. (1981). Tumour spectrum in the FAMMM syndrome. Br. J. Cancer 44: 553-560. MILLER, S. A., DYKES, D. D., AND POLESKY, H. F. (1988). A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res. 16: 1215. NAKAMURA, Y., LEPPERT, M., O’CONNELL, P., WOLFF, R., HOLM, T., CULVER, M., MARTIN, C., FUJIMOTO, E., HOFF, M., KUMLIN, E., AND WHITE, R. (1987). Variable number tandem repeat (VNTR) markers for human gene mapping. Science 235: 1616-1622. NEMER, M., SIROIS, D., AND DROUIN, J. (1986). XhoI polymorphism at the human pronatriodilatin (hPND) gene locus. Nucleic Acids Res. 14: 8696. Orr, J. (1976). A computer program for linkage analysis of general pedigree data. Am. J. Hum. Genet. 28: 528-529.
23.
RAMSAY, R. G., CHEN, P., IMMRAY, F. P., KIDSON, C., LAVIN, M. F., AND HOCKEY, A. (1982). Familial melanoma associated with dominant ultraviolet radiation sensitivity. Cancer Res. 42: 2909-2912.
24.
VAN HAERINGEN, A., BERGMAN, W., NELEN, M. R., VAN DER KOOIJ-MEIJS, E., HENDRIKSE, I., WLJNEN, J. T., KHAN, P. M., KLASEN, E. C., AND FFUNTS, R. R. (1989). Exclusion of the dysplastic naevus syndrome (DNS) locus from the short arm of chromosome 1 by linkage studies in Dutch families. Genomics 5: 61-64.
25.
WEISS, M. J., SPIELMAN, R. S., AND HARRIS, H. (1987). high-frequency RFLP at the human liver/bone/kidney-type alkaline phosphatase locus. Nucleic Acids Res. 15: 860.
A