Mutation Analysis of the BRCA1 and BRCA2 Genes | Genetics | Biology

In this article we will discuss about the mutation analysis of the BRCA1 and BRCA2 genes.

The BRCA Genes and Hereditary Breast and Ovarian Cancer:

During the last decade several genes responsible for autosomal dominant transmission of greatly increased risk of specific cancers were identified. Although it has been known for years that some families exhibited an appar­ently autosomal dominant pattern of early-onset breast and ovarian cancer, the characterization of the syndrome of hereditary breast and ovarian cancer (HBOC) predominantly resulted from the identification of the BRCA1 and BRCA2 genes and the ability to analyze the BRCA genes for mutations responsible for the HBOC syndrome.

Mutations in both genes are highly penetrant and con­fer an increased risk not only for HBOC, but also for other malignancies, e.g., colon and prostate cancer. Early comprehensive studies of the Breast Cancer Linkage Consortium assessed the contribution of BRCA1 and BRCA2 to inherit­ed breast cancer by linkage and mutation analysis in families with at least 4 cases of breast cancer.

In an estimated 52% of families disease was linked to BRCA1, in 32% of families to BRCA2, and to neither gene in 16% indicating other predisposing genes as well. In the majority of families (81%) with both cancers, breast and ovarian, cancer was linked to BRCA1. In most other cases cancer was linked to BRCA2. The majority of families (76%) with mate breast cancer cases were caused by mutations in the BRCA2 gene.

The largest propor­tion of BRCA1/BRCA2 negative families (67%) were families with female breast cancer only. Testing large series of families in a clinical setting, e.g., the large extended families in which the BRCA genes were originally discovered, with mul­tiple individuals who had developed breast or ovarian cancer at an early age, were found not to be typical of most BRCA families.

A recent study aimed at assessing characteristics of family history that correlate best with the presence of BRCA1 and BRCA2 mutations in individuals tested in a clinical setting. Mutations were identified in 17.2 % of the 10,000 individuals tested, including 968 (20%) of 4,843 women with breast cancer and 281 (34%) of 824 women with ovarian cancer. Mutation prevalence in high-risk women of European ancestry was 16% (712/4379).

This is comparable with the risk of women of other, non-Ashkenazi ancestries. Interestingly, 36% of individuals tested in this study reported a family history of only a single first-degree or second-degree relative with early-onset breast or ovarian cancer.

Among the latter, 13% (224/1,693) carried deleterious mutations indicating that family history of BRCA1 or BRCA2 mutation carriers ascertained in a clinical setting with a cer­tain amount of small family pedigrees may be neither dramatic nor obvious.

An essential prerequisite for achieving clinical benefits from the cloning of the BRCA genes is finding and assessing the significance of mutations in these susceptibility genes.

In order to coordinate the detection and interpretation of BRCA mutations and to make the results available to as many qualified investi­gators as possible, a central repository of information regarding mutations and polymorphisms, the Breast Cancer Information Core (BIC), was created and is being maintained by an international collaborative effort hosted by NHGRI.

A total of 6672 and 5624 sequence vari­ations/entries for BRCA1 and BRCA2, respectively, have been registered with the BIC database. These entries represent 1237 and 1381 distinct mutations, polymorphisms and variants of the BRCA1 and BRCA2 gene, respectively.

A total of 711 and 872 sequence variants for BRCA1 and BRCA2, respectively, were reported only once. Most of the distinct alterations reported are deleterious mutations such as frame shift, nonsense (574 for BRCA1 and 576 for BRCA2), or splice site mutations (132 for BRCA1 and 69 for BRCA2).

Implementation of informative BRCA1 and BRCA2 testing programs is aided by the acquisition of population-specific genetic data. In Europe, such a com­prehensive study with more than 400 individuals has been performed only for the BRCA1 gene in the Dutch population.

In other populations like the French, Swedish and the Finnish, 100-200 families have been screened for aberrations in the BRCA1 gene. In each of these populations less than 150 families have been investigated for mutations in the BRCA2 gene. The proportion of cases attributed to BRCA1 ranges from 10% in Finland and the Netherlands to approximately 25% in France and Sweden.

BRCA2 mutation detection rates are generally around 10%, with the exception of Italy where germline mutations have been found in as much as 25% of 49 families analyzed.

In the German population, where most studies have been limited to the analysis of BRCA1 or to relatively small series, a com­prehensive study was initiated in 1996 by the formation of a research collabo­ration, the “German Consortium for Hereditary Breast and Ovarian Cancer,” sup­ported by the Deutsche Krebshilfe (German Cancer Aid).

Organization and Aims of the German Consortium:

With the possibility of testing for alterations in the two BRCA genes known to predispose for breast and ovarian cancer, identification of individuals at risk and their inclusion in surveillance programs became feasible.

Careful evalua­tion of paternal as well as maternal family history is required, especially in women diagnosed with breast cancer before the age of 50 or ovarian cancer at any age, to enable the appropriate identification and counseling of individuals at risk for carrying mutations in BRCA1 and BRCA2. Breast and/or ovarian can­cer patients who belong to families with multiple cases of early onset breast and ovarian cancer are strong candidates for BRCA mutations.

BRCA testing is predictive because a given BRCA mutation does not inevitably lead to cancer in a given individual. A BRCA mutation markedly increases the likelihood of developing breast and ovarian cancer, but other genetic factors and environmental influences also contribute to the phenotype.

A BRCA mutation predicts a higher lifetime risk for cancer, but does not make the diagnosis a certainty. Similarly, a true negative result, e.g., the absence of a detectable mutation, for a member of a high-risk family predicts that the risk of cancer is not markedly increased over that of the average woman in the pop­ulation. Therefore, the Utter still have a probability of 10 to 12% of develop­ing sporadic breast cancer and 1-2% of sporadic ovarian cancer.

Testing patients at high risk in an established clinical trial requires stan­dardized procedures and confidentiality. The impact of breast and ovarian can­cers on patients and their families can be considerable. Exploring a patient’s emotional and psychological concerns before initiating tests is recommended. In addition, genetic counseling is not just for the individual being tested, but is directed at the entire family.

In extending consultation to interested family members the possible ethical-philosophical implications (informed consent, personal autonomy), psychosocial implications (perception of risk, emotional state), and legal implications must be kept in mind. Therefore, the complex­ities surrounding BRCA testing suggested an interdisciplinary counseling strat­egy.

Taking the aforementioned considerations into account, the German Cancer Aid decided to fund ten research centers in Germany starting in 1997. To ensure ease of access for families, centers were distributed all over Germany – Berlin, Bonn, Dusseldorf, Frankfurt, Kiel, Heidelberg, Munchen, Munster, Ulm, and Wurzburg. In the following years two additional centers in the eastern part of Germany – Dresden and Leipzig – were included.

All centers were committed to the same inclusion criteria and counseling strategies. The study protocol uses a multidisciplinary approach that provides counseling and genetic testing by bringing together, in each center, a team of genetic counselors, gynecologists and psycho-oncologists. Each local ethics committee approved the protocol used. Bilateral cancer or breast and ovarian cancer of the same patient were counted as independent cases of breast/ovarian cancer.

Inclusion criteria were based on family history as follows Families with 2 or more cases of breast can­cer, which included at least one case with onset under the age of 50 years; fam­ilies with one or more cases of breast and at least 1 case of ovarian cancer; fam­ilies with at least 1 male breast cancer, and families with a single case of breast cancer with age of diagnosis before 35 years.

Representatives of the three dis­ciplines performed counseling in either three consecutive sessions or in one ses­sion with all members present. Referrals to the study have come from univer­sity hospitals, private practice physicians, as well as by self-referral.

After a first counseling session, eligible candidates had a minimum of four weeks to consider participating in the study. After a second counseling session, blood was drawn for genetic testing. Test results were presented in a third counseling session.

In the beginning different strategies for the analysis of the BRCA genes were used in the individual centers. Some centers uses SSCP and/or PTT as pre- screening methods followed by DNA-sequencing of positive results. Others rou­tinely performed direct DNA-sequencing.

At the beginning the center in Kiel, in collaboration with Peter Oefner at Stanford University, tested the usefulness of the DHPLC technique as a prescreening method for BRCA analysis. After ini­tial, promising results, the center in Kiel initiated a comparative study between SSCP, DHPLC, and sequencing. The results were presented to the other cen­ters and convinced 10 of them to change their analysis strategy according to the Kiel procedure.

Established Techniques in (BRCA1/2) Mutation Analysis:

Mutation detection has found important applications in functional genomics and clinical diagnostics. Realizing the full benefits of the decoding of the human genome requires rapid and sensitive methods for screening genomic DNA for unknown sequence variations.

Technologies used for the lat­ter task should be capable of fully automated, high-throughput analysis that ideally does not require modified PCR primers, customized specific reagent arrays, detection labels, or any sample pre-treatment other than PCR.

Numerous techniques are available for mutation detection like direct sequencing (DS), protein truncation test (PTT), single-strand conformation polymorphism (SSCP), dideoxy fingerprinting assay (DDF), denaturing gradient gel elec­trophoresis (DGGE), two-dimensional gene scanning (TDGS), conforma­tion-sensitive get electrophoresis (CSGE), enzymatic mutation detection (EMD), allele-specific oligonucleotide hybridization (ASO), and immobi­lized DNA hybridization assays.

Up to now, none of these methods meet the challenges set forth above. The widely used SSCP analysis is economical and simple, but has low sensitivity, ranging from 60 to 90%. However, the major disadvantage of SSCP is the lack of its objective predictability. Thus, since there are no theoretical assurances, non-appearance of a band-shift does not prove the absence of a mutation.

By contrast, automated direct sequencing is high­ly sensitive but costly and labor intensive. DDF is a combination of a Sanger sequencing reaction with multiple-fragment SSCP and is more sensitive than SSCP alone, but still labor intensive. With PTT only sequence alterations lead­ing to a truncated protein can be detected. Other alterations (missense, inframe deletions, and insertions) escape the detection of the PTT assay.

Other methods, including DGGE and CSGE, have been used with some success, but despite their improved sensitivity over SSCP, these methods are technically challenging and formatted for manual use only. TDGS is a method for analyzing multiple DNA fragments in parallel for all possible sequence variations, using extensive multiplex PCR and two-dimensional electrophoretic separation on the basis of size and melting temperature.

One source of error is the inter­pretation of the complex spot patterns produced by this method. Also, the sen­sitivity of TD6S is impaired by frequent preferential amplification of the non- mutant allele, resulting in very “light” heteroduplices that obscured accurate reading of the gels.

EMD methods for mutation scanning still lack the sen­sitivity and specificity of the chemical cleavage of the mismatch method. Other approaches such as immobilized DNA hybridization arrays still have significant false positive signal, as well as high cost per assay.

Denaturing high performance liquid chromatography (DHPLC) has generat­ed increased interest over the past years in clinical genetics, because of its potential for automation and its ease of application. The technique, which is based on heteroduplex detection, allows for automated identification of single- nucleotide substitutions and small deletions or insertions. Heteroduplex pro­files are easily distinguished from homoduplex peaks.

Furthermore, DHPLC has been shown to clearly resolve mutations in various genes with detection rates ranging from 92.5 to 100%. An updated version of the list of genes screened entirely or partly by DHPLC is available at the web site of the Stanford Genome Technology Center, Palo Alto, USA.

Development of a DHPLC Based BRCA1/2 Gene Test in Routine Genetic Diagnostics:

The two BRCA genes together comprise approximately 15,700 nucleotides of open reading frame. More than 1,000 mutations of deduced or established clinical significance have been identified and are distributed throughout the 48 coding exons and respective splice junctions of the two genes. In families with both breast and ovarian cancer the detection rate for BRCA1 and BRCA2 muta­tions is only 43% and 10%, respectively.

The BRCA mutation frequencies in most families were even less because of the inclusion criteria for BC/OC fami­lies to be tested within the German study, which required one case of early- onset breast cancer (< 50 years) and one first relative with breast cancer at any age. Therefore, a selective prescreening to avoid labor and cost intensive sequencing of wild type exons is an indispensable necessity.

Number of Amplified BRCA1/BRCA2 Fragments:

Primer pairs for the amplification of BRCA1 and BRCA2 exons and intron- boundary regions were chosen with respect to the length and optimal DHPLC performance of the PCR products. Samples suitable for mutation detection by DHPLC are PCR products in the range 200 – 500 base pairs (bp).

Fragments larg­er than 500 bp can be run, but sensitivity for mutation detection may be reduced. Fragments of less than about 200 bp in length do not give satis­factory results, as the fragments melt over too narrow a temperature range.

For oligonucleotide design, software provided through the world wide web were utilized.

The underlying sequences were based on GenBank entries for BRCA1 (L78833) and BRCA2 (Z74739). As shown in Table 3-1 and Table 3-2, the BRCA1 gene was divided into 34 PCR fragments. Lengths of the fragments ranged from 234 to 564 bp.

The lengths of the 42 PCR fragments dividing the BRCA2 gene ranged from 278 to 707 bp. Some of the oligonucleotide primers were published earlier. The primer pairs described are those used by the majority of the cen­ters of the German consortium. PCR needs to be optimized, because poor qual­ity PCR products will produce poor quality mutation detection/DHPLC results.

Establishing Running Conditions:

Separation of homoduplex and heteroduplex DNA, and hence mutation detection, by DHPLC is temperature dependent. Selecting appropriate temper­atures for the analysis of each individual fragment is critical to the success of DHPLC. There are three different methods for predicting analysis temperatures. One possible method is the generation of an experimental melting curve for determining optimal analysis temperatures.

In order to generate an experi­mental melting curve, a wild type fragment is run at progressively increasing temperatures as shown in Figure 3-1A. The temperature steps (1°C) in the example given here for BRCA2, exon 4 cover a temperature range from 50 to 56°C. For the example presented here, a significant reduction in retention time is observed at each temperature step.

At temperatures above 56°C no further effect on retention time is observed under the gradient conditions employed here, because the DNA fragment analyzed is completely denatured.

On the basis of this data set, a melting curve can be constructed by plotting retention time as a function of analysis temperature with end points at 50°C, corresponding to 100 % double-strand character of the fragment analyzed and 56°C, correspon­ding to 100 % single-strand character of the fragment analyzed.

In this partic­ular case, the temperature point at which 50 % of the fragment analyzed was denatured was determined to be 53°C. The analysis temperature is generally set one to two degree below the experimentally determined temperature point at which 50 % of the fragment is denatured.

Figure 3-1B shows the elution pro­files of a positive control (BRCA2, IVS4+67A->C) and the detection range for this mutation (51-55°C). The “best” result, most clearly distinguishing/resolv­ing both hetero- and homoduplices is seen at 53°C.

This experimental approach of determining melting curves of wild type and respective positive controls is probably the single best approach by which the optimal analysis temperature can be determined. However, multiple melting domains may not always be apparent from the experimental melting curve and positive controls are often not available.

Programs for the theoretical prediction of melting behavior/profiles of DNA fragments provide useful additional information. WAVEMAKER software from Transgenomic, Inc. and the “Melt” program. HYPERLINK provide accurate and reliable prediction of melting temperature(s) for DNA fragments analyzed.

A representative exam­ple of a melt profile generated with WAVEMAKER software is given in Figure 3- 2. The 450-bp DNA fragment of BRCA2, exonl6 displays two melting domains. The majority of the DNA fragment is melting at around 54°C to 55°C, whereas a narrow region of about 30 bp remains double-stranded even at a temperature of as high as 59°C (“high-melting” domain).

Consequently, a mutation like the IVSl5-116 del TAG, which is located in the low-melting part of this DNA frag­ment, is detected at 54°C. In contrast, mutation G7986A, which is located in the high-melting domain, is seen only at 58°C. This example emphasizes the need for using theoretical melting curves predicted with WAVEMAKER software.

Furthermore, it points out the superior sensitivity of the DHPLC method com­pared to the TGGE, where the G7986A alteration could not be detected (Figure 3-2). In most cases the aforementioned melt programs provide reliable analy­sis conditions allowing one to avoid the laborious generation of experimental melting curves. If possible, heteroduplex resolution should be additionally checked with positive controls that harbor known sequence variations.

The group in Dusseldorf collected a large panel of samples of different BRCA1/BRCA2 mutations, which are listed in the BIC database, for the German BRCA Consortium by asking the depositors into the database for DNA samples or PCR products.

Due to the good feedback, the panel of positive controls includes at least one mutation for each BRCA1/BRCA2 fragment and in most cases for each melting domain. All positive controls were PCR amplified and PCR products were distributed to each center of the Consortium for establishing and validating DHPLC running conditions.

DHPLC analysis was carried out using a WAVE Nucleic Acid Fragment Analysis System equipped with a DNASep Cartridge. Crude PCR samples were injected after heteroduplex formation and eluted at a flow rate of 0.9 ml/min using a linear gradient of acetonitrile (ACN) in 0.1 M triethylammonium acetate (TEAA), pH 7.0. The software-predicted ACN gradient is adjusted in such a way that the elution time of the peaks falls with­in a 3- to 4-minute window.

Generally, the analysis of an individual sample takes 8-10 min. This includes regeneration and equilibration of the column after each analysis. A representative elution gradient is shown in Figure 3-3. DHPLC running conditions for all amplicons of the BRCAI and BRCA2 gene are summarized in Table 3-3 and Table 3-4.

The analysis portion of each gradient is preceded by a loading drop of 0.5 min, e.g.: the PCR product to be analyzed is loaded onto the DNASep Cartridge at a percent Buffer B that is 5 % lower than the starting percentage of Buffer B of the analysis portion of the gradient.

In order to maximize the sensitivity of the technology and detect all possible sequence variations, especially in fragments harboring different melting domains, it is recommended to perform mutation analysis at several different temperatures. In many cases more than one analysis temperature per amplicon is listed in tables 3 and 4.

If the temperatures given differ by only one or two degrees, the second analysis temperature is generally required in order to dis­tinguish between different sequence variants (see below) by DHPLCM. Therefore, a second run is necessary and is performed only in those cases that show a heteroduplex profile.

For example, the run at 49°C with the BRCA1/exon 2 fragment is to distinguish between the profiles produced by the intronic variation (IVSI-115t>c) alone and in combination with the intronic sequence vari­ant 185delAG. Analysis of the BRCAI/exon 111 fragment at two different tem­peratures permits differentiation between the sequence variants 3232A>G and 3238G>A based on their peak profiles.

Due to observed variations in column oven performance, which will affect the reproducibility of DHPLC peak pattern, the conditions provided may need adjustment if transferred from one instrument to another. Therefore, in routine genetic diagnostics it is recommended to check established running conditions with a relevant panel of positive controls before each run.

DHPLC is more than a Pre-Screening Technique:

DHPLC is typically used as a pre-screening technique, followed by DNA- sequencing for confirmation and identification of the sequence alteration lead­ing to a DHPLC-positive result. In the case that amplified gene regions contain known, common polymorphisms the sequencing costs can markedly increase.

If these harmless variations cannot be identified unambiguously, the advantage of a pre-screening method turns out to be of minor impact, because a large num­ber of samples have to be sequenced in the end. This is the case for the BRCA1/BRCA2 genes where some fragments are highly polymorphic.

In order to overcome this problem, two approaches have been followed:

(A) Elution profiles produced by individual sequence alterations, e.g., com­mon polymorphisms, have been identified and characterized.

(B) DHPLC-positive samples are re-analysis after mixing equal quantities of the DHPLC-positive sample with a sample containing the sequence alteration in question.

Some of the known BRCA1/BRCA2 variants produce specific chromatogram profiles that differ from known polymorphisms also occurring in the same fragments. In some cases the specific differences in elution profiles become most apparent if DHPLC analysis is conducted at different temperatures. Therefore, analysis of relevant DNA-fragments at different temperatures will increase repro­ducibility and reliability of characteristic mutant patterns (see above and Figure 3-4).

However, the reliability of a qualitative DHPLC analysis has to be tested for each gene individually because some sequence regions may produce very similar or even identical DHPLC peak patterns for different variants.

The impor­tance of multiple temperature analysis as well as frequent column cleaning is emphasized. Successful separation of heteroduplices on the DNASep cartridge is influenced by several factors. We observed that peak resolution might decrease with increasing age of the column.

Consequently, an objective assess­ment of column performance becomes necessary. This is of particular impor­tance for exons where few control samples are available that can be used to assess optimal column performance and differentiation between peak profiles. Several factors can influence column resolution, e.g., column batch, number of injections, additives in PCR reactions, or inadequate treatment of the column.

Daily cleaning of the column with 75% acetonitrile at 60°C prolongs column performance tremendously. We have obtained up to 7000 injections on a sin­gle column using the resolution of the mutation standards from Transgenomic as a determinant for column performance and resolution.

Nevertheless, it seems very difficult to judge if column performance is sufficient to predict specific sequence variations reliably. We propose to perform column performance testing at least once per week, which may include running several samples with similar elution profiles.

Hereby, the resolution of DHPLC peak patterns can be observed in consecutive runs. When test samples for column performance indi­cate altered heteroduplex resolution and cleaning does not improve the per­formance, the column should be changed.

A second alternative for confirming DHPLC-positive results of DNA frag­ments containing frequent polymorphisms is illustrated in Figure 3-5. Equal quantities of PCR product of unknown samples and a control DNA with a single known heterozygous sequence alteration are mixed, heated, and re-analyzed after renaturation using the same DHPLC conditions that were used to detect the sequence variation initially.

The absence of change in the elution profile of the mixed sample clearly indicates that the mutation in the sample is the same as in the control. Mutations in the sample other than those contained in the control lead to distinct patterns comparable with elution profiles of samples containing two heterozygous mutations.

This procedure requires the prepara­tion of a panel of control samples and an additional DHPLC analysis for each positive DNA-fragment. Using the above approach, mutations or polymorphisms can be reliably identified and DNA-sequencing can be omitted.

Sensitivity of DHPLC and Quality Management of BRCA1/2 Mutation Analysis in the German Consortium for Hereditary Breast and Ovarian Cancer:

The efficacy of DHPLC for mutational screening of the BRCA1 and BRCA2 genes is generally accepted. There are some papers in recent years document­ing the sensitivity and specificity of DHPLC in mutation detection. In a recent­ly published study comparing SSCP, CSGE, TDGS and DHPLC in BRCA1 mutation analysis, only the laboratory using DHPLC correctly identified all mutations.

Direct sequence analysis of 626 BRCA1 fragments previously subjected to DHPLC analysis confirmed the high sensitivity and specificity of DHPLC, both of which were 100%. In a more recent report, DHPLC detected all of 432 heterozygotes in BRCA1 and 136 in BRCA2, with only one false positive.

By comparing the chromatographic peak pattern of samples with that of standards containing known alternations, the identities of mutations were determined correctly in all but two cases without the need of sequencing.

In a separate study, out of 186 BRCA1 and BRCA2 mutations, only two sequence variations in exon 5 of BRCA1 yielded profiles difficult to distinguish from the chromatogram of a homozygous control. The reason may lie in this particular fragment’s ability to form a stable secondary cruciform structure, warranting its direct sequenc­ing.

But even if this and nine other BRCA1 amplicons that contain 16 frequent sequence variants were to be sequenced routinely to avoid missing mutations that yield DHPLC profiles similar to those of known polymorphisms, the cost saving obtained with DHPLC are significant due to the great number of amplicons (N – 34) required to cover the entire BRCA1 gene.

In conclusion, the sensitivity and specificity of DHPLC appear to be con­sistently higher than 96%. Not all failures can be readily explained. Challenging are mutations located in extremely high-melting domains sur­rounded by sequence that melts at significantly lower temperatures.

As mentioned above the German Consortium for Hereditary Breast and Ovarian Cancer started routine BRCA1/BRCA2 mutation analysis in 1997 and in 1999. Ten centers of the consortium established simultaneously the DHPLC methodology for germline mutation screening in the BRCA1/BRCA2 genes on a WAVE Nucleic Acid Fragment Analysis System.

In order to reach and maintain a high level of quality of genetic testing for BRCA genes the consortium members decided to implement a program for external quality assessment (EQA). For this purpose, an external expert in the field provided a BRCA1/BRCA2-specific EQA scheme.

DNA samples were sent out with mock clin­ical referrals and a request for diagnosis. Written reports, including genotyping results and interpretations, had to be returned. The aim of this evaluation was to measure the actual, average performance level and to estimate the quality of testing performed by individual laboratories.

A set of 9 genomic DNA samples with BRCA1/BRCA2 mutations was sent out and 9 exons for each sample had to be analyzed. Results, including sequence alteration identified and their classification, had to be reported. In total 410 BRCA1/BRCA2 amplicons were analyzed. In 5 cases (1.22%) false negative DHPLC results were obtained. The error rate for the identification and classifi­cation of DHPLC positive results was 1,73 %.

Each false negative DHPLC result occurred only once per exon and the mutation was detected in the other 8 lab­oratories correctly. Moreover 4/5 errors occurred in 2 centers indicating that DHPLC sensitivity in general is above 98%. We conclude that the accuracy and sensitivity of DHPLC for mutation detection is in part dependent on the skill level and experience in interpreting elution profiles of the person(s) performing the analysis.

While an accuracy of over 98 % may appear good at first glance, it must be kept in mind that much of the BRCA1/BRCA2 testing is done as pre­dictive testing on individuals who may have an increased family risk but are free of the cancer disease. Clearly, predictive testing needs the highest possible reli­ability.

Results from the German Consortium for Hereditary Breast and Ovarian Cancer:

The main focus of this German-wide, multi-center study was to establish a BRCA1/2 mutation profile and to determine family types with high frequencies of mutations in these genes. In a comprehensive study, the entire coding sequences of the breast cancer genes BRCA1 and BRCA2 were analyzed in 989 unrelated patients from German breast/ovarian cancer families.

A total of 77 BRCA1 and 63 BRCA2 distinct deleterious mutations were found in 302 patients. More than 1/3 of these mutations are novel and might be specific to the German population. Eighteen common mutations were found in 68% of cases in BRCA1 and 13 recurrent mutations in 44% of cases in BRCA2, facilitating prescreening approaches. Haplotype analyses indicate that 14 out of 20 recurrent mutations are likely originating from a common founder.

An additional 50 different rare sequence variants with unknown relevance for tumor genesis were found in 72 families. Correlation of BRCA1/BRCA2 detec­tion rates with family history identified families with both breast and ovarian cancer to be at highest risk for BRCA1/BRCA2 mutations (43% and 10%, respec­tively), followed by families with at least 2 pre-menopausal cases of breast can­cer (24% BRCA1 and 13% BRCA2 mutations).

These data provide strong evi­dence for further predisposing genes in the German population. In breast can­cer families with 2 or 3 affected females but only a single or no pre-menopausal case, mutations were detected with low frequencies (about 10% or less for both genes). The decision for or against molecular diagnosis is now aided by con­sidering the expected mutation detection rates that greatly depend on family history and structure.