Determination of Hereditary Mutations in the BRCA1 Gene Using

Departments of Surgical Oncology, Pathology, Gynecological Oncology, Central Laboratory, and Tumor Biology, Institute for Cancer Research, The Norwegi...
1 downloads 0 Views 55KB Size
Anal. Chem. 2004, 76, 4406-4409

Determination of Hereditary Mutations in the BRCA1 Gene Using Archived Serum Samples and Capillary Electrophoresis Per O. Ekstrøm,† Tone Bjørge,‡ Anne Dørum,§ Ane Sager Longva,| Karen-Marie Heintz,| David J. Warren,⊥ Svein Hansen,# Randi Elin Gislefoss,O and Eivind Hovig*,|

Departments of Surgical Oncology, Pathology, Gynecological Oncology, Central Laboratory, and Tumor Biology, Institute for Cancer Research, The Norwegian Radium Hospital, Oslo, Norway, and Institute of Population-Based Cancer Research, The Cancer Registry of Norway, Institute of Clinical Biochemistry, Rikshospitalet, Oslo, Norway

Analysis of DNA variation in biological samples most frequently utilizes the polymerase chain reaction (PCR) performed on extracted genomic DNA, followed by visualization of alleles using various methodologies. Few reports have demonstrated that amplification of DNA from plasma and serum samples is possible. We have performed DNA amplification on a large set of serum samples (n ) 2955). Here, we report that known hereditary mutations in the BRCA gene can efficiently be analyzed in serum samples collected and stored over several decades. Fragments were PCR-amplified following a short initial denaturation of the serum sample in a standard microwave oven. Fragment analysis was subsequently performed using a DNA capillary-sequencing instrument. The PCR success rates were fragment- and size-dependent ranging from 83.2% to 97.9%. Of the 11 820 polymerase chain reactions performed, the overall PCR success rate was 91.3% (10 796/11 820), which is comparable to PCR performed on genomic DNA. The advantage of the method described herein is its ability to utilize archival material stored in serum biobanks for long periods of time. The nearly completed human genome sequence provides the reference against which all other human sequencing data can be compared. Following the massive sequencing effort by the human genome project1 and the privately funded approach by Celera,2 the search for links between DNA variations (mutations or singlenucleotide polymorphisms) and phenotypes causing disease has escalated. * To whom correspondence should be addressed. Phone: +47 22 93 54 16. Fax: + 47 22 52 24 21. E-mail: [email protected]. † Department of Surgical Oncology, The Norwegian Radium Hospital. ‡ Department of Pathology, The Norwegian Radium Hospital. § Department of Gynecological Oncology, The Norwegian Radium Hospital. | Department of Tumor Biology, The Norwegian Radium Hospital. ⊥ Department of Central Laboratory, The Norwegian Radium Hospital. # Institute of Population-Based Cancer Research. O Institute of Clinical Biochemistry. (1) Collins, F. S.; Morgan, M.; Patrinos, A. Science 2003, 300, 286-90. (2) Adams, M. D.; Sutton, G. G.; Smith, H. O.; Myers, E. W.; Venter, J. C. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 3025-26.

4406 Analytical Chemistry, Vol. 76, No. 15, August 1, 2004

A variety of biological materials have been used as sources of genomic DNA for PCR in such studies, the most common being leukocytes or whole blood. Unfortunately, clinical laboratories have not generally stored such samples; instead, many laboratories have accumulated large banks of frozen sera, which could turn out to be a useful source of DNA for genetic studies. In the context of a case-control study, we aimed at analyzing a large number of serum samples. The results from the case-control study will be published elsewhere. When using serum as a source of DNA, several technical aspects needed to be addressed. Depending on the sample size available, DNA purification from serum is feasible. However, DNA extraction is labor-intensive, time-consuming, and costly and poses a potential risk of crosscontamination. Automated DNA purification methods have been aimed at preventing some of these problems, but they increase the need for expensive equipment and consumables. In serum, DNA is present in low concentration and is usually highly fragmented; thus, the number of PCR cycles has to be increased and the length of target sequences should preferably be kept below 100 base pairs. Furthermore, with limiting amounts of serum, sample handing, that is, the number of pipetting steps, becomes important due to possible loss of DNA. We prepared serum samples with regard to the above principles and analyzed 2955 samples for the four most common inherited mutations in the breast/ovarian cancer gene BRCA1 reported for the Norwegian population.3 With a simple initial denaturation step by microwave irradiation, followed by the PCR amplification of target sequences and visualization by capillary electrophoresis, a total of 11 820 PCR reactions on serum gave an overall success rate of 91.3%. MATERIALS AND METHODS Samples. We received 30 µL of serum from 2955 samples from the Janus serum depository.4,5 These samples were analyzed for the four most common germ line mutations in the BRCA1 gene, as reported for the Norwegian population (1675 del A, 1135 ins (3) Borg, A.; Dørum, A.; Heimdal, K.; Mæhle, L.; Hovig, E.; Møller, P. Dis. Markers 1999, 15, 79-84. (4) Jellum, E.; Andersen, A.; Lundlarsen, P.; Theodorsen, L.; Orjasæter, H. Environ. Health Perspect. 1995, 103, 85-88. (5) Jellum, E.; Andersen, A.; Lundlarsen, P.; Theodorsen, L.; Orjasæter, H. Sci. Total Environ. 1993, 140, 527-35. 10.1021/ac049788k CCC: $27.50

© 2004 American Chemical Society Published on Web 06/12/2004

Table 1. Primers Used for Amplification on Serum, Product Size, Annealing Temperatures, and Number of PCR Cycles primer ID and direction BRCA1 1675 del A forward BRCA1 1675 del A reverse BRCA1 1135 ins A forward BRCA1 1135 ins A reverse BRCA1 3347 del AG forward BRCA1 3347 del AG reverse BRCA1 816 del GT forward BRCA1 816 del GT reverse

annealing temp

cycles

5′6FAM-TAC-ATC-AGG-CCT-TCA-TCC-T-3′ 64 base pair 5′-AGG-AGT-CTT-TTG-AAC-TGC-3′

60 °C

40

5′-TGG-GCT-GGA-AGT-AAG-GA-3′ 90 base pair 5′6FAM-TCT-CTC-ACA-CAG-GGG-ATC-AG-3′

58 °C

42

5′-ACA-TTC-AAG-CAG-AAC-TAG-GT-3′ 62 base pair 5′6FAM-CCC-CTA-ATC-TAA-GCA-TAG-CA-3′

58 °C

42

5′-AGG-CTG-CTT-GTG-AAT-T-3′ 57 base pair 5′6FAM-TGG-GTT-GAT-GAT-GTT-CAG-TA-3′

55 °C

43

primer sequence

A, 3347 del AG, and 816 del GT).3 Aliquots of 5 µL of serum were transferred into 96 well microplates (Axygen, Tamro Medlab AS, Oslo, Norway). Serum aliquots were boiled in a microwave oven for 4 min at full power (1200 W). Then 40 µL of PCR master mix was added to each well, mixed, and subjected to temperature cycling as described below. Control Samples. Eight wells were reserved on each microplate and used for controls, two mutant positive, two mutant negative, and four wells without template. Primers. Primers used for the analysis are shown in Table 1. In each primer set, one of the primers was labeled with 6-carboxyfluorescein (6FAM). All primers were obtained from MedProbe (MedProbe, Oslo, Norway). Polymerase Chain Reaction (PCR). The final PCR reactions mixture consisted of approximately 3 µL of microwaved serum, 4 µL of 10X Buffer (Applied Biosystems, Oslo, Norway), 2.5 mM MgCl2, and 0.35 µL of 10 mM dNTP mix (AmershamBiosciences, Uppsala, Sweden). Primers were added in a final concentration of 0.15 mM of each primer. A combination of 1.5U Taq and 0.13U Pfu polymerase was used per well. Amplification was performed in an air-thermocycler (Peltier Thermal Cycler PTC 200, MJ Research, Waltham, MA), using the following cycling conditions: 5 min at 95 °C, X cycles of 20 s at 95 °C, 20 s at Y °C, and 30 s at 72 °C, where X and Y are described in Table 1 for each fragment. In all cases, the last cycle was followed by 6 min at 72 °C. Instrumentation. An unmodified capillary DNA sequencing instrument, MegaBACE 1000 (AmershamBiosciences, Uppsala, Sweden) was used for determination of fragment lengths. Samples were injected electrokinetically from the PCR plate, without any post-PCR cleanup, by applying a field of 145 V/cm for 30 s. Fragments were electrophoresed at 60 °C with a field of 145 V/cm for 30 min. DNA Concentration and Fragment Length Distribution in Serum. Selected serum samples were subjected to DNA extraction by standard phenol/chloroform procedures. Because of the high protein concentration in the samples, DNA extraction was performed in two steps. First, 150 µL of sterile water was added to 50-µL aliquots of serum. Then 200 µL of phenol-chloroform (1:1) was added followed by vigorous mixing and centrifugation at 10 000g. Next the upper phase (≈200 µL) from two tubes was

pooled followed by addition of 200 µL of phenol-chloroform (1: 1) and centrifuged as described above. After ethanol precipitation, DNA was dissolved in 5 µL of sterile water and analyzed using a DNA 7500 Assay on a Bioanalyzer 2100 (Agilent Technologies, Palo Alto, CA) for quantification of DNA and determination of size distribution of DNA fragments found in serum. RESULTS AND DISCUSSION Retrospective genetic studies depend on possible sources of DNA to be able to conduct molecular analysis of target sequences. Because subjects enrolled in the study may have passed away or are outside practical reach for the researcher, cell repositories or biobanks become important sources of biological material. In Norway, an ongoing serum bank, comprising approximately 700 000 samples consolidated from about 330 000 donors,6 was established in 1973 with the aim of elucidating relationships between sera biochemistry and risk/development of cancer. Because of the small sample volume (30 µL) available in the present study, anticipation of low DNA copy number, and expectancy of the yield, DNA extraction using standard protocols was considered inappropriate. Additionally, the cost of the extraction could not have been performed within the economical limits of the project. It has previously been demonstrated that adding 7 µL of serum into a PCR solution leads to complete inhibition of the reaction.7 However, it has been reported that microwave irradiation of serum efficiently denatures PCR inhibitors present in serum, while leaving the DNA amplifiable.8 This was tested by PCR amplification of aliquots of the 48 samples with and without microwave treatment. Omitting the microwave step resulted in no PCR products. To estimate the DNA sizes present in serum, we extracted DNA by standard phenol extraction procedures, followed by alcohol precipitation. The material obtained from several samples was pooled and analyzed on an Agilent Bioanalyzer for DNA size distribution. On the basis of the estimated DNA concentration in 100 µL of serum and assuming that the loss of DNA during extraction is negligible, 1 µL of serum contains (6) http://www.kreft.no/forhelsepersonell/janusserumbank/english/, (Accessed 2/6/04). (7) Ulvik, A.; Ueland, P. M. Clin. Chem. 2001, 47, 2050-53. (8) Sandford, A. J.; Pare, P. D. Biotechniques 1997, 23, 890-92.

Analytical Chemistry, Vol. 76, No. 15, August 1, 2004

4407

Figure 1. DNA fragment size and relative concentration estimated according to a known DNA ladder visualized on a DNA 7500 Assay on a Bioanalyzer 2100. DNA sizes in base pairs are indicated for the four peaks observed in the enlarged inset.

Figure 3. Twelve different samples amplified for fragment 3347 in different plates on different days and analyzed in independent runs. Sample one and seven are mutated while the remainders are wild type.

approximately 100 pg of DNA, which corresponds to 30 relatively fragmented genomic copies. Thus, the starting number of alleles in the PCR reaction would be in the range of 150 copies. This is in agreement with previous reports on the amount of DNA found in serum.9 Additionally, assuming a PCR efficiency of 1.65, the number of DNA copies after amplification would be close to 1010 per microliter, within the range for visualization by laser-induced fluorescence. As shown in Figure 1, a pattern consistent with an apoptotic DNA ladder could be observed, with very small amounts having a size above 377 base pairs. This is a reflection of the fact that the DNA in serum is to a significant extent the result of DNA being shed from apoptotic leukocytes.8 As expected, primers designed for consistent amplification using purified full length genomic DNA did not perform well, and

iterative selection of primers for smaller PCR product sizes was performed. We found that reproducible amplification was obtained when the primers were designed to produce products with lengths smaller than approximately 100 bp. Using optimized pre-PCR sample treatment, PCR design, and reaction parameters, we achieved success rates of 97.9%, 89.5%, 94.8%, and 83.2% for fragments 1675, 1135, 3347, and 816, respectively (fragment description, Table 1). The somewhat lower success rate of fragment 816 is fragment-specific since a similar low amplification rate was observed when purified genomic DNA was used as a template (data not shown). Because the inherited mutations analyzed were frame shift mutations, a simple fragment length analysis could be performed. Such analysis may be performed either with use of gel electrophoresis or by capillary electrophoresis. The latter allows higher sample throughput and the possibility of automated scoring by computer software. Figure 2 shows mutated and wild-type samples for each fragment analyzed. Noteworthy is that fragment 816 and 1135 do not have nontemplate addition of adenine, while fragment 1675 and 3347 have additional stutter peaks. Several factors may influence the formation of the nontemplate adenine additions, such as cycling conditions, concentration of dATP, sequence of the three last bases in the strand copied, and enzyme proofreading

(9) Takayama, T.; Yamada, S.; Watanabe, Y.; Hirata, K.; Nagai, A.; Nakamura, I.; Bunai, Y.; Ohya, I. Leg. Med. 2001, 3, 109-13.

(10) Brownstein, M. J.; Carpten, J. D.; Smith, J. R. Biotechniques 1996, 20, 10048.

Figure 2. Eight representative electropherograms of mutated and wild-type sample for the four fragments analyzed. The upper electropherogram displays mutated genotypes in all four boxes, with the corresponding wild type shown below. Please note that the nontemplate addition of adenine by the polymerase does not interfere with the interpretation of the electropherograms. Also note the different time scales for the fragments.

4408 Analytical Chemistry, Vol. 76, No. 15, August 1, 2004

capabilities.10,11 Nonetheless, stutter peaks did not influence recognition of mutated samples. All identified mutants were verified by sequencing, while randomly selected wild-type samples gave normal sequence results (n ) 148); hence, the sensitivity of the assay was 100%. Figure 3 demonstrates the reproducibility of 12 different samples amplified and electrophoresed on different days. The electropherograms are not corrected for differences in current; thus, some variation in migration time is to be expected. Nevertheless, the peak pattern is highly reproducible and samples could be genotyped with appropriate recognition software. The method described herein is well-suited for retrospective studies, where genomic DNA is not available through fresh tissue or blood samples, and fragment analysis for screening purposes is generally limited to known inherited mutations. To widen the (11) Magnuson, V. L.; Ally, D. S.; Nylund, S. J.; Karanjawala, Z. E.; Rayman, J. B.; Knapp, J. I.; Lowe, A. L.; Ghosh, S.; Collins, F. S. Biotechniques 1996, 21, 700-09.

application range of the approach described, we are currently investigating the possibility of applying a high-throughput capillary version of denaturing gradient gel electrophoresis (DGGE) to serum samples for detection of unknown inherited mutations in genes of interest. In this study we have demonstrated that serum can serve as a source of DNA and can be reliably implemented in large-scale analysis of genetic variation. The cost of analysis with this approach is reduced to a minimum (cost of PCR). This may prove beneficial for many large-scale future retrospective studies. ACKNOWLEDGMENT This work received financial support from Torsteds and Grete Harbitz’ legacy, the Norwegian Foundation for Health and Rehabilitation through the Norwegian Cancer Society. Received for review February 6, 2004. Accepted April 28, 2004. AC049788K

Analytical Chemistry, Vol. 76, No. 15, August 1, 2004

4409