Skip Navigation


Human Molecular Genetics Advance Access originally published online on July 22, 2003
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow An erratum has been published
Right arrow All Versions of this Article:
12/18/2311    most recent
ddg245v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (27)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by de Jong, M. M.
Right arrow Articles by van der Graaf, W. T. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by de Jong, M. M.
Right arrow Articles by van der Graaf, W. T. A.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Human Molecular Genetics, 2003, Vol. 12, No. 18 2311-2319
DOI: 10.1093/hmg/ddg245
© 2003 Oxford University Press

The HLA class III subregion is responsible for an increased breast cancer risk

Mirjam M. de Jong1,2,3, Ilja M. Nolte2,4, Elisabeth G. E. de Vries1, Michael Schaapveld6, Jan H. Kleibeuker3, Elvira Oosterom4, Jan C. Oosterwijk2, Annemarie H. van der Hout2, Gerrit van der Steege4, Marcel Bruinenberg4, H. Marike Boezen5, Gerard J. te Meerman2 and Winette T. A. van der Graaf1,*

1Department of Medical Oncology, 2Department of Medical Genetics, 3Department of Gastroenterology, 4Department of Medical Biology and 5Department of Epidemiology, University Medical Center, Groningen, The Netherlands and 6Comprehensive Cancer Center Northern Netherlands, Groningen, The Netherlands

Received April 17, 2003; Accepted July 10, 2003


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 RESULTS
 DISCUSSION
 SUBJECTS AND METHODS
 REFERENCES
 
BRCA1 and BRCA2 germline mutations account for <5% of breast cancer cases. Less penetrant breast cancer susceptibility genes are likely to exist. Earlier studies have suggested involvement of the HLA region. The HLA region was genotyped with 24 microsatellite markers and markers for two single nucleotide polymorphisms (SNPs) in TNF{alpha} and TNFß, in germline DNA from 956 breast cancer patients and 1271 family-based controls. Association analyses and the haplotype sharing statistic (HSS) were used to search for differences in haplotype sharing between patients and controls. Based on criteria known to influence genetic breast cancer risk, patients were divided into groups of high, moderate and low risk. The HSS revealed a significant difference in mean haplotype sharing between patients and controls for four consecutive markers (D6S2671, TNFa, D6S2672 and MICA), the highest being at D6S2671 (P=0.017). Subgroup analyses showed that moderate-risk patients were responsible for this difference, with the strongest association for D6S2672 (P=0.0009). A single haplotype was more frequent and longer in moderate-risk patients than in controls. The results were confirmed with association analyses. Individuals homozygous for haplotype 110–184 (D6S2672-MICA) were observed in 9.0% of moderate-risk patients and 1.5% of controls [odds ratio (OR)=7.14], while heterozygotes were at a lower risk (OR=1.41), suggesting a recessive effect. No association was observed between the two SNPs in TNF{alpha} (-308) and TNFß (intron 1) and breast cancer risk. The results reveal a potential role of the HLA class III subregion in susceptibility to breast cancer in patients at moderate familial risk.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
RESULTS
 DISCUSSION
 SUBJECTS AND METHODS
 REFERENCES
 
There is increasing evidence that breast cancer is a heterogeneous disease, both phenotypically and genotypically (1). Hereditary factors are important in breast cancer susceptibility. Since BRCA1 and BRCA2 gene mutations do not explain the occurrence of breast cancer in most breast cancer-prone families, other breast cancer susceptibility genes likely exist, such as CHEK2 (2–4).

Segregation analyses have shown that the best fitting model for breast cancer families without BRCA1 and BRCA2 mutations, is a polygenic one (i.e. low-penetrance genes with multiplicative effects on risk). However, a model with a single recessive allele produces a similar fit. With both models, the frequency of the mutations is high (24–32%) (5–8). Hence, there is much interest in the search for low penetrant genes/variants for breast cancer, which exist with high prevalence in the general population. The population attributable risk (PAR) of breast cancer which can be explained by these genes might be considerably higher than the PAR caused by the rather rare mutations in the BRCA1 and BRCA2 genes (9–11).

The 3.6 Mb human leukocyte antigen (HLA) region is divided into three regions (centromeric to telomeric) coding for class II (1.1 Mb), class III (0.7 Mb), and class I (1.8 Mb) antigens, respectively (12). Several studies have examined the HLA class I region in sporadic breast cancer patients, mostly with phenotype-based methods. Most studies were relatively small and had, therefore, insufficient power to reach significance (13–17). Three studies examined the HLA class II region (18–20). Two of these studies revealed an association with a decreased breast cancer risk for DRB1*11 carriers (19,20). Two studies examined the HLA class III region (21,22). One Tunisian study examined the -308 SNP in TNF{alpha} and a SNP at position 1267 of the HSP70-2 (21). Homozygosity for the variant allele of both SNPs showed an increased breast cancer risk (TNF{alpha} SNP OR=4.4 and HSP70-2 SNP OR=7.2). One Asian study examined several SNPs in the promoter region of TNF{alpha} (i.e. the -1031, the -863, the -857, and the -308 SNP) and one SNP in intron 1 of TNFß (22). Breast cancer was not associated with SNPs in the TNF{alpha} promoter. However, homozygosity for the variant allele of the TNFß SNP, showed an increased breast cancer risk (OR=5.33).

The present study uses markers covering the entire HLA region to analyze in a large sample of breast cancer patients and family-based controls the involvement of the HLA region in breast cancer susceptibility.

Patients were assigned to high, moderate or low genetic risk groups, based on age at onset, bilaterality of breast cancer, number of first-degree relatives with breast cancer, co-existence of ovarian cancer or male breast cancer in the family (9,23–27).

Allele, genotype and two-locus haplotype association methods were used for the analyses. In addition the haplotype sharing statistic (HSS) was used, as this extracts extra information from phase and single marker tests as compared with association analysis (28,29).


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
DISCUSSION
 SUBJECTS AND METHODS
 REFERENCES
 
A total of 956 patients and 1271 family-based controls were included between July 1998 and January 2002, with 91 patients in the high-risk, 212 in the moderate-risk, and 653 in the low-risk group. The median age at diagnosis was 53 years (range 23–88). One marker (D6S2703) proved not to be in Hardy–Weinberg equilibrium and was excluded from the analyses. The percentage of unknown alleles (including inheritance errors) was 8.9% for all 23 markers combined.

HSS
Strong linkage disequilibrium (LD) was observed in the sample of patients as well as in the sample of controls over the whole region studied (P<0.001 for all pairs of consecutive markers, i.e. no extremer values than the observed D' values were seen in the 1000 randomized datasets). The HSS analysis revealed a difference in mean haplotype sharing between all patients and controls for D6S2671 (P=0.017), TNFa (P=0.022), D6S2672 (P=0.020), and MICA (P=0.042; Fig. 1). Subgroup analyses showed that the moderate-risk patients in particular were responsible for this difference. A difference in mean haplotype sharing between moderate-risk patients and controls was observed for seven consecutive markers [D6S2670 (P=0.017), D6S273 (P=0.0054), D6S2671 (P=0.002), TNFa (P=0.0017), D6S2672 (P=0.0009), MICA (P=0.0017), and D6S2673 (P=0.0043)], with D6S2672 most strongly associated with breast cancer. At this locus, moderate-risk patients shared on average 0.43 marker intervals towards the telomere and/or the centromere as compared with 0.28 marker intervals among all patients and 0.21 marker intervals among controls. Analyses in the low- and high-risk patients showed no differences in mean haplotype sharing versus controls. When the patients were divided solely based on the age at diagnosis (>50 and <50 or >45 and <45) or solely on number of relatives with breast cancer, no association with breast cancer was observed for these different subgroups (data not shown). Data on menopausal state were not available.



View larger version (18K):
[in this window]
[in a new window]
  		Figure 1. The HSS for all patients and for the different risk groups. On the x-axis, distance to the first marker (i.e. 6SL001) is indicated and on the y-axis the -log10 value of the significance of the difference in mean haplotype sharing between patients and controls.

 
A color-transformed graphical representation of haplotypes of moderate-risk patients and corresponding family-based controls is shown in Figure 2. This reveals one haplotype that is over-represented in patients versus controls (the red-colored block in Fig. 2). The core of this haplotype is formed by markers D6S2665, D6S2670, D6S273, D6S2671, TNFa, D6S2672, MICA, D6S2673, D6S2678 and D6S2694, notably the combination of alleles 247–200–136–161–120–110–184–192–307–315, respectively. Another preserved haplotype (blue-colored in Fig. 2) is as frequent among patients as among controls.



View larger version (57K):
[in this window]
[in a new window]
  		Figure 2. Graphical representation of haplotypes with the patients in the left half of the figure and the corresponding family-based controls in the right half. Each horizontal colored line represents one haplotype. On the x-axis, marker loci are indicated. Marker alleles are mapped to colors, white being a missing or phase-unknown allele. Different alleles at a locus have different colors. The choice of colors is such that only minimal changes in color appear from one marker locus to the next in haplotypes that are shared over a long distance. In the direction of the y-axis, haplotypes are clustered for largest similarity around marker D6S2672. This marker was chosen because the difference in mean haplotype sharing between patients and controls was maximal at this locus. The same clustering algorithm was applied to patient and control haplotypes. Only patient and control haplotypes with at least four phase-known alleles are displayed. The haplotype causing the difference between patients and controls as evaluated by HSS is the haplotype depicted as a red block in the middle of the figure.

 
A conditional HSS analysis of only haplotypes carrying the alleles of the most associated two-locus haplotype (see below, in the description of the results of the association analyses), i.e. 110–184 at D6S2672-MICA, reveals that the risk haplotype as mentioned before extends over a longer region among moderate-risk patients (on average, 2.9 marker intervals) than among all patients (on average, 2.6 marker intervals) and than among controls (on average, 2.2 marker intervals). This means that patients, in particular moderate-risk patients, have a more recent founder for this region than controls, which supports the use of the HSS technique as it indeed appears to extract additional information from the data over association analyses.

Association analysis
The results observed with the HSS were confirmed by association analysis. Table 1 presents the results of the allele, genotype and two-locus haplotype association analyses for all patients and for the different subgroups (P-values corrected for 69 independent tests). Table 2 displays the corresponding ORs, CIs and PARs.


View this table:
[in this window]
[in a new window]
  		Table 1. Frequencies of alleles, genotypes and two-locus haplotypes observed to cause an increased risk among moderate risk patients. The significant result is bold
 

View this table:
[in this window]
[in a new window]
  		Table 2. Odds ratios (ORs), 99.93%a confidence intervals (CIs) and (population) attributable risks [(P)ARs] for the alleles concerned, for all patients and for the moderate-risk patients for alleles, genotypes and two-locus haplotypes. Significant results are bold
 
In the moderate-risk group, breast cancer was associated with a two-locus haplotype, i.e. D6S2672 (allele 110) and MICA (allele 184). A single haplotype could be constructed from D6S2665 to D6S2694 [D6S2665 (allele 247)–D6S2670 (allele 200)–D6S273 (allele 136)–D6S2671 (allele 161)–TNFa (allele 120)–D6S2672 (allele 110)–MICA (allele 184)–D6S2673 (allele 192)–D6S2678 (allele 307)–D6S2694 (allele 315)]. This haplotype is the same haplotype as the one retrieved with HSS. Individuals homozygous for the D6S2672-MICA haplotype 110–184 were observed in 9.0% of moderate-risk patients and 1.5% of controls [odds ratio (OR)=7.14], while individuals heterozygous for the haplotype 110–184 were at a lower risk (OR=1.41).

The SNPs TNFA and TNFB were analyzed with allele, genotype and two-locus association analyses. The results of the genotype analysis are described in Table 3. The percentage of unknown alleles (including inheritance errors) was 6.5% for TNFA and 5.3% for TNFB. No association was observed between these SNPs and breast cancer. The wild-type alleles of both SNPs were present on the risk haplotype identified with HSS and association analyses.


View this table:
[in this window]
[in a new window]
  		Table 3. Frequencies of the genotypes of the TNFA and TNFB SNPs
 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
SUBJECTS AND METHODS
 REFERENCES
 
This is the largest study so far investigating the HLA region in germline DNA for breast cancer risk. It also is the first study of the HLA region performing subgroup analyses for breast cancer genetic risk with stratification based on age at onset, bilateral breast cancer, the occurrence of multiple first-degree relatives with breast cancer, male breast cancer and ovarian cancer.

The HLA class III subregion appeared to be significantly associated with breast cancer in all patients. Moderate-risk patients in particular were responsible for the association (attributable risk in moderate-risk patients 19.7%).

A single preserved haplotype was more frequent and longer among moderate-risk patients, as compared with controls. This haplotype extends from D6S2665 to D6S2694, as illustrated in the color-representation of the haplotypes (Fig. 2). This implies that the risk haplotype is not only more frequent among patients than among controls, but that the moderate-risk patients are also more related to each other than the entire patient population. Therefore, both hypotheses of HSS are met, resulting in more significants by HSS than by association analyses. An important implication of this, in a methological sense, is that haplotype analyses can extract more information from genetic data than association analyses alone. Homozygosity for the D6S2672–MICA haplotype had a stronger association with breast cancer than heterozygosity, suggesting a recessive effect. This is compatible with results of segregation analyses, where a model with a single recessive allele, with a high frequency, was one of the best fitting models for breast cancer families without BRCA1 and BRCA2 mutations (5–8). We detected only one haplotype associated with an increased breast cancer risk. The responsible sequence variant putatively present on this haplotype is expected to occur frequently in the population.

Genes relevant for breast cancer located in the region of D6S2671, TNFa, D6S2672, and MICA are TNF{alpha} and TNFß. The two SNPs analyzed in this study [-308 in TNF{alpha} (TNFA) and intron 1 in TNFß (TNFB)] were, however, not associated with breast cancer in this study.

In 243 Tunisian breast cancer patients and 174 controls (21), homozygosity for the variant allele of the -308 TNF{alpha} SNP, showed an increased risk (OR=4.4). However, a study in 95 Asian breast cancer patients and 190 controls revealed no association between breast cancer and several SNPs in the promoter region of TNF{alpha}, including the -308 SNP (22), although homozygosity for the variant allele of the TNFß SNP showed an increased breast cancer risk in the second study.

Our data revealed that our risk haplotype contained the wild-type alleles of both SNPs analyzed in this study. This implies that the SNPs themselves are probably not responsible for the increased breast cancer risk. However, the variant responsible for the increased breast cancer risk may well be the same in the three populations, as it could be in LD with the variant -308 TNF{alpha} allele in the Tunisian population, with the variant TNFß allele in the Asian population and with the wild-type alleles of both SNPs in our population.

It should be noted that, apart from the aforementioned genes, many other genes are located in the region between loci D6S2671–MICA. Based on what is known of the function of these genes at this moment, their involvement in breast cancer susceptibility is unlikely.

The fact that we have investigated a large sample of breast cancer patients and controls, and performed subgroup analyses for three genetic risk groups may well be the reason why, in contrast to previous studies with inconclusive results, we observe a strong association between breast cancer and the HLA region.

The putative breast cancer susceptibility gene(s) will probably give a lower breast cancer risk than the well-known mutations in BRCA1 and BRCA2, since the effect is observed mainly in the moderate-risk group. It seems difficult to reconcile the finding of an effect in the moderate-risk patients but not the high-risk patients, however, the high-risk patients are expected to be explained in a higher percentage by BRCA1 and BRCA2 mutations. Furthermore, less penetrant genes are supposed to be more frequent in smaller familial clusters (30).

The risk haplotype was frequent among moderate-risk patients (24.9%). Therefore, the variant present on this haplotype is also expected to be frequent. In the moderate-risk group, the attributable risk was high, namely 23.4%. Because of the high frequency of the observed risk haplotype, the PAR of the putative breast cancer causing variant is expected to be considerably higher than the PAR caused by mutations in the BRCA1 and BRCA2 genes (the PAR of BRCA1 and BRCA2 combined is expected to be less than 5%) (31,32).

Recently, the cell-cycle checkpoint kinase CHEK2 gene was identified as a low-penetrance breast cancer susceptibility gene (4,33). It was estimated that the 1100delC mutation of CHEK2 resulted in an approximately 2-fold increased breast cancer risk in women and therefore the demand for clinical testing of such an allele is questionable (4). In our study, the OR is 7.14 for the association with markers D6S2672 and MICA in moderate-risk patients. When the underlying variant is identified, this increased breast cancer risk may well require clinical testing. With an OR of seven, the putative breast cancer gene is not really low-penetrant and moderate-penetrant seems to be a more appropriate term.

It will be of great importance to further study this region with the aim to find the gene and mutations involved. This may not only lead to the identification of individuals at increased risk, but may also provide new insights into the etiology of breast cancer (34). With a frequent haplotype, several SNPs located in the region are expected to be in LD with the mutation and it will be laborious to find the causal SNP or variant based on association analyses only. In the ideal situation, all SNPs located in the candidate region should be genotyped. It is important to realize, however, that there are already 290 validated SNPs in our 500 kb candidate region, 170 of them positively tested for heterozygosity. Analysis of (part of) these SNPs in our patient and control sample is one of our future goals.


   SUBJECTS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 SUBJECTS AND METHODS
REFERENCES
 
Patients and controls
Both patients and family-based controls participated in a population-based study set-up to detect breast cancer susceptibility genes. These individuals were part of the Caucasian population of the northern Netherlands. Patients were accrued from six hospitals (the University Medical Center in Groningen, the Medical Centers in Leeuwarden and Harlingen, Ny Smellinghe in Drachten, Talma Sionsberg in Dokkum and the Anthonius Hospital in Sneek). In the hospitals in Drachten and Dokkum, breast cancer patients were identified through the registry of the Comprehensive Cancer Center Northern Netherlands. In the other four hospitals, all breast cancer patients visiting the outpatient clinic over a 1 year period were asked to participate. Their physician in the outpatient clinic always approached the breast cancer patients. No selection was performed; cases were included between July 1998 and January 2002, irrespective of family history. Patients completed a questionnaire including their family history of cancer. Only those patients were excluded in whom the presence of a BRCA1 or BRCA2 mutation was known. Twenty-two patients in the study were known to have undergone mutation analyses and tested negative for BRCA1 and BRCA2 (11 in the high-risk group, four in the moderate-risk group and seven in the low-risk group). Since only the index cases were matched, it is possible that in family members mutation analyses have been performed in Groningen or that patients or family members have undergone mutation analyses at another genetic center in the Netherlands.

All DNA samples and data in this study were handled anonymously and individuals were aware that they would not be informed about individual test results. The Medical Ethical Committees of the participating hospitals approved the study. All included subjects gave written informed consent.

Family members (preferably parents or child and spouse) served as controls. When DNA from parents was available for phase determination of the alleles, the non-transmitted haplotype from each parent was used as control. If DNA was obtained from a child and a spouse, both haplotypes of the spouse were regarded as controls. When only one family member (a parent, child or sib) was available, the haplotype not present in the patient was used as a control haplotype.

Subgroup analyses were performed for patients of high, moderate and low genetic risk, respectively, based on age at onset, bilaterality of breast cancer and family history of breast cancer, ovarian cancer and/or male breast cancer (criteria for the different subgroups are presented in Table 4). In the present study, tests for BRCA1 and BRCA2 mutations were not performed.


View this table:
[in this window]
[in a new window]
  		Table 4. Criteria based on family history as used in this paper to define the different genetic risk groups for breast cancer susceptibility (9,23–27)
 
Genotyping
DNA was extracted from 20 ml EDTA-blood following standard procedures and was stored at -80°C. Twenty-four polymorphic microsatellite markers in the HLA region at 6p21 were used for genotyping. All markers are located within the HLA region, in particular within the HLA class II region. The marker order was determined by using sequence data as published by the MHC sequencing consortium (see Table 5 and Fig. 3 for marker information) (35). For the genes located in this region, the NCBI map (www.ncbi.nlm.nih.gov/cgi-bin/Entrez/map-search) and the Celera maps (Celera Genomics, Rockville, MA, USA) were used. The volume of PCR reactions was 10 µl, which included ~25 ng DNA. For each polymerase chain reaction (PCR), 0.5 units Taq DNA polymerase (Amersham Pharmacia Biotech, Uppsala, Sweden) were used to amplify the fragments. Reaction mixtures contained 0.2 mM dNTP (Roche Diagnostics, Mannheim, Germany), 2.5 mM MgCl2, 10 mM Tris–HCl (pH 9.0), 50 mM KCl (Amersham Pharmacia Biotech) and 0.25 µM of each primer (with one primer 5' labeled with a fluorochrome 6-FAM, HEX [(Sigma, Malden, the Netherlands) or NED (Applied Biosystems, Foster City, CA, USA)]. Cycling was performed on a PTC-225 thermal cycler (MJ Research, Waltham, MA, USA) and a PrimusHT (MWG Biotech, Ebersberg, Germany). A standard protocol was used for amplification. Post PCR multiplexing was performed by combining 2–10 µl (based on signal strength) of the PCR products. A 2.3 µl sample of the pooled fragments was mixed with 2.5 µl MilliQ and 0.2 µl ET-400R size standard (Amersham Pharmacia Biotech) and separated on a MegaBACE 1000 capillary sequencer (Amersham Pharmacia Biotech) according to the manufacturers protocol. Results were analyzed using genetic profiler version 1.1 (Amersham Pharmacia Biotech). Scoring of the alleles was blinded for affection status and family structure.


View this table:
[in this window]
[in a new window]
  		Table 5. Marker data and primer sequences
 


View larger version (22K):
[in this window]
[in a new window]
  		Figure 3. Map of markers and genes in the region constructed according to the published sequence of the HLA region (25) and other available sources (the NCBI and the Celera map). The positions of the markers used for analysis are indicated above the bar and the genes (in italics) below the bar. The distances indicated on the scale are in megabases.

 
In addition, two SNPs were analyzed, one in TNF{alpha} gene (TNFA, -308) and one in TNFß (TNFB, intron 1, NcoI RFLP). The sequences containing information of the direct surrounding of the SNPs were retrieved from GenBank (TNFA, AF279459; and TNFB, AY070490), and the 300 bp around the SNPs was screened for additional variations by BLAST search (www.ncbi.nlm.nih.gov/BLAST/). The additional variations found were masked with an N, so no additional variants are present in the primers and probes. The resulting sequences were used for submission to the Assay-by-DesignTM (Applied Biosystems). The following primers and probes were used: TNFA-308 forward primer, GAAATGGAGGCAATAGGTTTTGAG; TNFA reverse primer, GGCCACTGACTGATTTGTGTGTAG; TNFA probe variant allele (T), VIC-CCGTCCTCATGCC-MGBNFQ; TNFA probe wild-type allele (C), 6FAM-CCGTCCCCATGCC-MGBNFQ; TNFB forward primer, CAGTCTCATTGTCTCTGTCACACATT; TNFB reverse primer, ACAGAGAGAGACAGGAAGGGAACA; TNFB probe variant allele (G), 6FAM-CCATGGTTCCTCTC-MGBNFQ; TNFB probe wild-type allele (A), VIC-CTGCCATGATTCC-MGBNFQ. Reactions were performed in 5 µl volumes and contained 25 ng DNA, 1xTaqMan Universal PCR Master Mix (Applied Biosystems), 100 nM of each primer and 900 nM of each probe. Cycling conditions on the ABI prism 7900 HT (Applied Biosystems) were 2 min 50°C, 10 min 95°C followed by 40 cycles of 15 s 92°C and 1 min 60°C. End-point fluorescence was measured immediately after cycling. Alleles were assigned using SDS 2.0 software (Applied Biosystems).

Statistical methods
After genotyping all individuals with the 24 markers, the set of haplotypes present in patients and the set of the control haplotypes were determined. As mentioned above, the non-transmitted haplotypes of the parents or the haplotypes of the spouse of the patient served as control haplotypes. In case of missing data, all available family members (i.e. parents, sibs or children of the patients) were used (in an attempt) to reconstruct the missing data.

Differences between these two haplotype sets were analyzed using the HSS, which analyses the length of haplotype similarity. The validity of this method has been demonstrated previously both in simulation studies and empirical data (28,29,36). The HSS assumes that haplotype segments of patients from a founder population are conserved in the region spanning a disease locus. In contrast, control individuals are not expected to have conserved haplotypes centered at a particular locus. In a region with strong LD, a well-defined founder population is less important for the analyses compared to a region with less strong LD.

To evaluate conservation of haplotypes, similarity or sharing between a pair of haplotype is measured by counting, from a marker locus both in centromeric and telomeric direction, the number of consecutive markers that have identical alleles on both haplotypes. The haplotype sharing is calculated for each marker locus, which, as haplotype sharing extends over multiple marker loci, produces highly correlated values, and consequently dependent tests for the difference between patients and controls. This haplotype sharing is next computed for all pairs of haplotypes, separately for the patient haplotypes and the control ones. The hypothesis of HSS is that patients will display an excess of haplotype sharing, as compared to controls, which is expected to be maximal at the marker loci closest to the susceptibility or disease locus. When haplotype sharing is smaller among patient haplotypes than among control ones, this is indicated by a negative -log10(P-value) in the HSS plot. This might then be evidence for a protective disease susceptibility locus, but more research is necessary in that area.

As a test of LD in order to prove the presence of conserved haplotypes, the D' for multi-allelic markers was used (37). The significance of the observed D'-value was assessed by randomization. For each locus, the observed alleles were randomly distributed over the haplotypes. The significance of the observed D'-value was determined by the fraction of 1000 randomizations that revealed a larger D'-value than the observed one.

In addition to the HSS, allele, genotype and two-locus haplotype association analyses were performed. For these analyses, the frequencies of the different alleles, genotypes and two-locus haplotypes, respectively, were compared between patients and controls using a t-test to assess significant differences. This t-test assumes that the patients and controls are independent samples from the population and that the number of a specific allele, genotype or two-locus haplotype follows a binomial distribution that can be approximated by a normal distribution (significant associations are defined by a corrected P-value below 0.05). Strength of association was determined based on the HSS results, i.e. only the alleles, genotypes and two-locus haplotypes were analyzed which were identified with HSS to be associated with breast cancer.

For the subgroup analyses for familial risk of breast cancer, each subgroup was compared with the entire control group.

A multiple testing correction was performed for 23 markers in three subgroups using a Bonferoni correction, implying that reported P-values and confidence intervals (CIs) are corrected for 69 independent tests. It should be noted that this is a very conservative approach, in particular for the haplotype analysis, as strong LD exists over the entire region under study. ORs and their 95% CIs were calculated from the observed number of patients and controls without adjusting for any external variables.

In order to give an indication of the PAR of the causal mutation, PARs of microsatellite alleles were determined for all patients. Similarly, attributable risks (ARs) for the moderate-risk patients were determined, since for this category the effect was most pronounced.

For allele and genotype associations, all patients were included. For the HSS and two-locus haplotype association analyses, which require phase to be derived, only those with at least one family member were used.

In case of missing or phase unknown alleles, haplotype sharing was averaged over all possible allele combinations weighted by the a priori probabilities. For an unknown allele, the a priori probability was set equal to the allele frequency and for a phase unknown allele it was set to 0.5, thus information on LD between markers was not used in this procedure. The actual software for the HSS test and for the phase derivation as well as the genotyping data presented here will be available by an online link via www.med.rug.nl.


   ACKNOWLEDGEMENTS
 
We thank all families who donated their DNA to this project. We would also like to thank the co-workers of the Departments of Surgery and Medical Oncology (University Medical Center Groningen, Medical Center Leeuwarden and Harlingen, Ny Smellinghe Drachten, Talma Sionsberg Dokkum, and Anthonius Hospital Sneek) and of Radiotherapy (University Medical Center Groningen). Furthermore we would like to thank Giny Bokma, Gerry Sieling and the blood bank Sanquin Noord-Oost (location Groningen) for their work in respectively patient and blood collection. We thank Drs C.H.C.M. Buys and D.S. Postma for fruitful comments. This work was supported by grant RUG-98-1665 of the Dutch Cancer Society and by a grant from the Comprehensive Cancer Center Northern Netherlands.


   FOOTNOTES
 
* To whom correspondence should be addressed at: Department of Medical Oncology, University Medical Center Groningen, Hanzeplein 1, PO Box 30.001, 9700 RB, Groningen, The Netherlands. Tel: +31 503612821; Fax: +31 503614862; Email: w.t.a.van.der.graaf{at}int.azg.nl Back


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 SUBJECTS AND METHODS
 REFERENCES
 

  1. Aubele, M. and Werner, M. (1999) Heterogeneity in breast cancer and the problem of relevance of findings. Anal. Cell Pathol., 19, 53–58.[Web of Science][Medline]

  2. Phelan, C.M., Lancaster, J.M., Tonin, P., Gumbs, C., Cochran, C., Carter, R., Ghadirian, P., Perret, C., Moslehi, R., Dion, F. et al. (1996) Mutation analysis of the BRCA2 gene in 49 site-specific breast cancer families. Nat. Genet., 13, 120–122.[CrossRef][Web of Science][Medline]

  3. Easton, D. (1997) Breast cancer genes—what are the real risks? Nat. Genet., 16, 210–211.[CrossRef][Web of Science][Medline]

  4. Meijers-Heijboer, H., van den Ouweland, A., Klijn, J., Wasielewski, M., de Snoo, A., Oldenburg, R., Hollestelle, A., Houben, M., Crepin, E., van Veghel-Plandsoen, M. et al. (2002) Low-penetrance susceptibility to breast cancer due to CHEK2(*)1100delC in noncarriers of BRCA1 or BRCA2 mutations. Nat. Genet., 31, 55–59.[CrossRef][Web of Science][Medline]

  5. Easton, D.F. (1999) How many more breast cancer predisposition genes are there? Breast Cancer Res., 1, 14–17.[CrossRef][Medline]

  6. Cui, J., Antoniou, A.C., Dite, G.S., Southey, M.C., Venter, D.J., Easton, D.F., Giles, G.G., McCredie, M.R. and Hopper, J.L. (2001) After BRCA1 and BRCA2-what next? Multifactorial segregation analyses of three-generation, population-based Australian families affected by female breast cancer. Am. J. Hum. Genet., 68, 420–431.[CrossRef][Web of Science][Medline]

  7. Antoniou, A.C., Pharoah, P.D., McMullan, G., Day, N.E., Ponder, B.A. and Easton, D. (2001) Evidence for further breast cancer susceptibility genes in addition to BRCA1 and BRCA2 in a population-based study. Genet. Epidemiol., 21, 1–18.[CrossRef][Web of Science][Medline]

  8. Antoniou, A.C., Pharoah, P.D., McMullan, G., Day, N.E., Stratton, M.R., Peto, J., Ponder, B.J. and Easton, D.F. (2002) A comprehensive model for familial breast cancer incorporating BRCA1, BRCA2 and other genes. Br. J. Cancer, 86, 76–83.[CrossRef][Web of Science][Medline]

  9. Rebbeck, T.R. (1999) Inherited genetic predisposition in breast cancer. A population-based perspective. Cancer, 86, 2493–2501.[CrossRef][Web of Science][Medline]

  10. te Meerman, G.J. and de Vries, E.G.E. (2001) Relevance of high and low penetrance. Lancet, 358, 331–332.[Web of Science][Medline]

  11. de Jong, M.M., Nolte, I.M., te Meerman, G.J., van der Graaf, W.T., Oosterwijk, J.C., Kleibeuker, J.H., Schaapveld, M. and de Vries, E.G. (2002) Genes other than BRCA1 and BRCA2 involved in breast cancer susceptibility. J. Med. Genet., 39, 225–242.[Abstract/Free Full Text]

  12. Tsujimura, A., Ota, M., Katsuyama, Y., Sada, M., Miura, H., Matsumiya, K., Gotoh, R., Nakatani, T., Okuyama, A. and Takahara, S. (2002) Susceptibility gene for non-obstructive azoospermia located near HLA-DR and -DQ loci in the HLA class II region. Hum. Genet., 110, 192–197.[CrossRef][Web of Science][Medline]

  13. Bouillenne, C. and Deneufbourg, J.M. (1979) Positive correlation between breast cancer incidence and HLA antigens. Oncology, 36, 156–159.[Web of Science][Medline]

  14. Hammond, M.G., Appadoo, B. and Brain, P. (1979) HLA and cancer in South African indians. Tissue Antigens, 14, 296–302.[Web of Science][Medline]

  15. Subira, M.L., Crisci, C.D., Zornoza, G., Sanz, M.L., Voltas, J., Hernandez, J.L. and Oehling, A. (1979) Breast cancer and histocompatibility antigens. Allergol. Immunopathol. (Madr.), 7, 411–416.[Medline]

  16. Abrahamova, J. and Majsky, A. (1988) HLA system and some neoplastic diseases. Acta Univ. Carol. Med. Monogr., 123, 1–80.[Medline]

  17. Biswal, B.M., Kumar, R., Julka, P.K., Sharma, U. and Vaidya, M.C. (1998) Human leucocytic antigens (HLA) in breast cancer. Ind. J. Med. Sci., 52, 177–183.[Medline]

  18. Casoli, C., Zanelli, P., Adorni, A., Starcich, B.R. and Neri, T. (1994) Serological and molecular study on the HLA phenotype of female breast cancer patients. Eur. J. Cancer, 30A, 1207–1208.[CrossRef][Web of Science][Medline]

  19. Chaudhuri, S., Cariappa, A., Tang, M., Bell, D., Haber, D.A., Isselbacher, K.J., Finkelstein, D., Forcione, D. and Pillai, S. (2000) Genetic susceptibility to breast cancer: HLA DQB*03032 and HLA DRB1*11 may represent protective alleles. Proc. Natl Acad. Sci. USA, 97, 11451–11454.[Abstract/Free Full Text]

  20. Ghaderi, A., Talei, A., Gharesi-Fard, B., Farjadian, S.H., Amirzargar, A. and Vasei, M. (2001) HLA-DBR 1 alleles and the susceptibility of Iranian patients with breast cancer. Pathol. Oncol. Res., 7, 39–41.[Medline]

  21. Mestiri, S., Bouaouina, N., Ahmed, S.B., Khedhaier, A., Jrad, B.B., Remadi, S. and Chouchane, L. (2001) Genetic variation in the tumor necrosis factor-alpha promoter region and in the stress protein hsp70-2: susceptibility and prognostic implications in breast carcinoma. Cancer, 91, 672–678.[CrossRef][Web of Science][Medline]

  22. Park, K.S., Mok, J.W., Ko, H.E., Tokunaga, K. and Lee, M.H. (2002) Polymorphisms of tumour necrosis factors A and B in breast cancer. Eur. J. Immunogenet., 29, 7–10.[CrossRef][Web of Science][Medline]

  23. Evans, D.G., Fentiman, I.S., McPherson, K., Asbury, D., Ponder, B.A. and Howell, A. (1994) Familial breast cancer. Br. Med. J., 308, 183–187.[Abstract/Free Full Text]

  24. Pharoah, P.D., Day, N.E., Duffy, S., Easton, D.F. and Ponder, B.A. (1997) Family history and the risk of breast cancer: a systematic review and meta-analysis. Int. J. Cancer, 71, 800–809.[CrossRef][Web of Science][Medline]

  25. Ford, D., Easton, D.F., Stratton, M., Narod, S., Goldgar, D., Devilee, P., Bishop, D.T., Weber, B., Lenoir, G., Chang, C.J. et al. (1998) Genetic heterogeneity and penetrance analysis of the BRCA1 and BRCA2 genes in breast cancer families. Am. J. Hum. Genet., 62, 676–689.[CrossRef][Web of Science][Medline]

  26. Peto, J., Collins, N., Barfoot, R., Seal, S., Warren, W., Rahman, N., Easton, D.F., Evans, C., Deacon, J. and Stratton, M.R. (1999) Prevalence of BRCA1 and BRCA2 gene mutations in patients with early-onset breast cancer. J. Natl Cancer Inst., 91, 943–949.[Abstract/Free Full Text]

  27. Nathanson, K.L. and Weber, B.L. (2001) ‘Other’ breast cancer susceptibility genes: searching for more holy grail. Hum. Mol. Genet., 10, 715–720.[Abstract/Free Full Text]

  28. Beckmann, L., Fischer, C., Deck, K.G., Nolte, I.M., te Meerman, G.J. and Chang-Claude, J. (2001) Exploring haplotype sharing methods in general and isolated populations to detect gene(s) of a complex genetic trait. Genet. Epidemiol., 21 (Suppl. 1), S554–S559.

  29. Levinson, D.F., Kirby, A., Slepner, S., Nolte, I., Spijker, G.T. and te Meerman, G.J. (2001) Simulation studies of detection of a complex disease in a partially isolated population. Am. J. Med. Genet., 105, 65–70.[CrossRef][Web of Science][Medline]

  30. Rahman, N., and Stratton, M.R. (1998) The genetics of breast cancer susceptibility. A. Rev. Genet., 32, 95–121.[CrossRef][Web of Science][Medline]

  31. Hopper, J.L., Southey, M.C., Dite, G.S., Jolley, D.J., Giles, G.G., McCredie, M.R., Easton, D.F. and Venter, D.J. (1999) Population-based estimate of the average age-specific cumulative risk of breast cancer for a defined set of protein-truncating mutations in BRCA1 and BRCA2. Australian Breast Cancer Family Study. Cancer Epidemiol. Biomarkers Prev., 8, 741–747.[Abstract/Free Full Text]

  32. Anglian Breast Cancer Study Group (2000) Prevalence and penetrance of BRCA1 and BRCA2 mutations in a population-based series of breast cancer cases. Br. J. Cancer, 83, 1301–1308.[CrossRef][Web of Science][Medline]

  33. Vahteristo, P., Bartkova, J., Eerola, H., Syrjakoski, K., Ojala, S., Kilpivaara, O., Tamminen, A., Kononen, J., Aittomaki, K., Heikkila, P. et al. (2002) A CHEK2 genetic variant contributing to a substantial fraction of familial breast cancer. Am. J. Hum. Genet., 71, 432–438.[CrossRef][Web of Science][Medline]

  34. Pritchard, J.K. and Cox, N.J. (2002) The allelic architecture of human disease genes: common disease-common variant ... or not? Hum. Mol. Genet., 11, 2417–2423.[Abstract/Free Full Text]

  35. The MHC Sequencing Consortium (1999) Complete sequence and gene map of a human major histocompatibility complex. Nature, 401, 921–923.[CrossRef][Medline]

  36. Boon, M., Nolte, I.M., Bruinenberg, M., Spijker, G.T., Terpstra, P., Raelson, J., De Keyser, J., Zwanikken, C.P., Hulsbeek, M., Hofstra, R.M., Buys, C.H. and te Meerman, G.J. (2001) Mapping of a susceptibility gene for multiple sclerosis to the 51 kb interval between G511525 and D6S1666 using a new method of haplotype sharing analysis. Neurogenetics, 3, 221–230.[Web of Science][Medline]

  37. Lewontin, R.C. (1964) The interaction of selection and linkage. I. General considerations; heterotic models. Genetics, 49, 49–67.[Free Full Text]


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 J. Med. Genet.Home page
M Zoodsma, I M Nolte, M Schipper, E Oosterom, G van der Steege, E G E de Vries, G J te Meerman, and A G J van der Zee
Analysis of the entire HLA region in susceptibility for cervical cancer: a comprehensive study
J. Med. Genet., August 1, 2005; 42(8): e49 - e49.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow An erratum has been published
Right arrow All Versions of this Article:
12/18/2311    most recent
ddg245v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (27)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by de Jong, M. M.
Right arrow Articles by van der Graaf, W. T. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by de Jong, M. M.
Right arrow Articles by van der Graaf, W. T. A.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?