| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1 Institut für Biochemie, Emil-Fischer-Zentrum, and 2 Pathologisch-Anatomisches Institut, Friedrich-Alexander Universität Erlangen-Nürnberg, Erlangen, Germany
aaddress correspondence to this author at: Institut für Biochemie, Emil-Fischer-Zentrum, Universität Erlangen-Nürnberg, Fahrstrasse 17, D-91054 Erlangen, Germany; fax 49-9131-85-22485, e-mail katrin.schiebel{at}biochem.uni-erlangen.de
A variety of genetic alterations are associated with the initial steps of carcinogenesis in sporadic tumors. Inactivation of tumor suppressor genes frequently occurs in a sequential process of genomic deletion of one allele and missense or nonsense mutation of the other allele in somatic cells (1). Cytogenetically undetectable deletions can be identified at the molecular level as loss of heterozygosity (LOH). Comparative genotyping of polymorphic markers such as microsatellites or single-nucleotide polymorphisms (SNPs) in healthy and tumor tissue can detect the loss of one allele by demonstrating the conversion of a heterozygous marker to a hemizygous genotype. LOH analysis of solid tumors does not necessarily show 100% deletion of one allele, as blood and immune cells without LOH may contaminate the tumor. Therefore, LOH analysis often shows a reduction, rather than a complete disappearance, of one allele.
Genotyping of polymorphic markers has been performed predominantly by gel-based methods, using microsatellites or short tandem repeats. Because of the repetitive nature of these polymorphic structures, PCR amplification often leads to so-called "shadow bands" resulting from DNA polymerase slippage. Shadow bands may overlap with major polymorphic bands and are frequently sources of genotyping ambiguities, particularly in the case of apparent incomplete deletion of one allele (2).
Breast cancer is the most common malignancy of women, with up to 95% of cases sporadic (3). Cytogenetic investigations and microsatellite LOH analysis have revealed deletions in the terminal part of Xp in breast carcinoma (4)(5). On the basis of analysis of 13 SNPs within the genomic region of the potential tumor suppressor gene PPP2R3B, located in the terminal band Xp22.3 (6)(7), we established a matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS) approach to analyze LOH in nonmicrodissected tissue.
Comparative sequencing of PCR-amplified fragments in tumor and healthy tissue revealed reduced allelic signals in tumor tissue in 3 of 29 patients, whereas in healthy tissue, sequencing signals of both alleles were within 20% of each other (Fig. 1 in the Data Supplement that accompanies the online version of this Technical Brief athttp://www.clinchem.org/content/vol51/issue3/). All three (patients 2, 5, and 20) showed loss of several heterozygous markers, indicating that true genomic deletion had occurred.
This study was approved by the Ethics Committee of the University Erlangen-Nürnberg for the Department of Gynaecology and Obstetrics, and informed consent was obtained from all participating patients. We used proteinase K digestion and standard protocols to extract DNA from frozen samples of histologically separated tumor and the surrounding healthy tissue of breast cancer patients treated at the Department of Gynecology. PCR reactions (50 µL) contained 25 pmol of each primer (PPPIE2for, 5'-GACTCTTCAGGTTCAGACACGG-3'; PPPE2Irev, 5'-CACGTACCTTGGCCACCAGG; PPPIE10/11for, 5'-CCGCAGGGAAGATCAG-3'; PPPIE12rev, 5'-TCACCCGTCCTCCCAGG-3'; PPPI12Cfor, 5'-ATTCGTGGAGCCGGCGTC-3'; PPPI13Crev, 5'-CAGAGGGTGTCCGTGTGG-3'; all from MWG Biotech), 50 ng of DNA template, 1 U of Taq polymerase in the supplied PCR buffer (Promega) supplemented with 1.5 mM MgCl2, and 2 nmol of each deoxynucleotide triphosphate. PCR conditions were 96 °C for 2 min; 35 cycles of 96 °C for 30 s, 30 s at 65 °C (exon 2; 210 bp) or 60 °C (exon 11–12; 300 bp), and 72 °C for 30 s; followed by 5 min at 72 °C. Exon 13 (370 bp) was amplified by touchdown PCR with an annealing temperature of 64 °C for 10 cycles and 56 °C for 30 cycles.
Purified PCR products (GenoPure ds; Bruker Daltonics) were used as templates for 10-µL extension reactions containing 1–2 U of Thermosequenase, 1x reaction buffer (Amersham Biosciences), 2 nmol of assay-specific deoxynucleotide triphosphates and dideoxynucleotide triphosphates, and 12 pmol of extension primer. For example, primer PPPE11C40678Tfor (5'-TGAAGCGCTGCAAGCTGGC-3') was used with a mixture of dTTP, ddCTP, and ddATP with extension at 58 °C (see Table 1 in the online Data Supplement). Extension reactions were performed at 96 °C for 2 min, followed by 40 cycles at 94 °C for 30 s, 53–62 °C for 30 s, and finally 72 °C for 3 min. We analyzed 1-µL aliquots of purified extension samples (GenoPure oligo; Bruker Daltonics) by MALDI-TOF-MS (Bruker Daltonics) according to the method described by Bonk et al. (8).
To quantify the loss of a specific allele, we analyzed DNA samples from healthy and corresponding tumor tissue in parallel, using shared master mixtures. We compared the peak-height ratio obtained by MALDI-TOF-MS of the healthy tissue with that of the tumor tissue, as described previously (8), and analyzed the results by use of an unpaired t-test.
All seven DNA fragments showing allelic reduction by sequencing were confirmed by MALDI-TOF-MS. Shown in Fig. 1A
are the MALDI-TOF-MS spectra at nucleotide 40678 in exon 11 for samples from patients 2, 20, and 7. Whereas the peak heights in healthy tissue from patients 2 and 20 were very similar, the peak heights for the T allele in tumor tissue were lower. In contrast, samples from patient 7, who did not exhibit LOH, gave similar peak heights for both healthy and tumor tissue.
|
We analyzed DNA extracted from healthy and tumor breast tissue of five individuals by repeated primary PCR reactions and in up to seven parallel extension reactions to exclude random variation and to assess the reproducibility of LOH measurements. Two patients with LOH (2 and 20) and three patients (7, 10, and 22), characterized by sequencing as lacking LOH, were reanalyzed by MALDI-TOF-MS for SNP C40678T. In the healthy tissue, the ratio of mean peak heights (T/C alleles) ranged between 0.76 and 0.92 but did not differ significantly in patients with (2 and 20) or without (7, 10, and 22) LOH (Table 1
).
|
Analysis of tumor tissue in non-LOH patients (7, 10, and 22) also revealed mean peak ratios in the range of 0.78–0.94 with no significant difference between healthy and tumor tissue (P >0.05). In contrast, patients for whom LOH had been shown by sequencing had significantly different peak ratios in tumor and healthy tissues (Table 1
). Repeated analysis of the exon 11 polymorphism in DNA from tumor and nontumor samples from patients 2, 7, 10, 20, and 22 clearly demonstrated that the tumor tissue of patients 2 and 20 had lost significant amounts of the T allele, giving mean peak ratios of 0.20–0.29. No such allelic reduction was observed in patients 7, 10, and 22, for whom signal ratios in tumor and healthy tissue were in the same range (Fig. 1B
and Table 1
).
In patient 2, signal ratios in tumor tissue for three additional heterozygous markers, located in intron 11 (A40768G and T40788G) and exon 12 (C40907T), were reduced for all SNPs analyzed, indicating that the whole region was deleted in the tumor (Fig. 1C
). Similarly, one allele of exon 11, intron 11 (A+B), and exon 12 also was reduced in patient 20.
To characterize the ability of MALDI-TOF-MS to detect LOH, we mixed (in ratios of 1:10, 1:1, and 10:1) two synthetic alleles, oligonucleotides PPPE11tempG (5'-GACGTTGGCCAGCTTGCAGCGCTTCAGG-3') and PPPE11tempA (5'-GACGTTAGCCAGCTTGCAGCGCTTCAGG-3'), representing the exon 11 polymorphism C40678T. The signal height ratio was linearly related to the ratio of template DNA variants used in the extension reactions. Whereas the same amount of PPPE11tempG and PPPE11tempA gave a signal ratio of 1.04, diverging ratios of template DNA led to significantly different peak-height ratios. Only marginal differences in signal ratios were detected for multiple samples set up in parallel (Fig. 2 in the online Data Supplement).
Comparative genomic hybridization was developed to identify large sections (10–20 Mbp) of increased or deleted genomic material in tumors. Array comparative genomic hybridization with spotted bacterial artificial chromosomes or with ligation-mediated PCR of bacterial artificial chromosomes can narrow the intervals for reliable copy number measurements to 
1.4 Mbp (9), whereas the method of multiplex ligation-dependent probe amplification (MLPA) (10) detects mid-size deletions. Use of MLPA in cancer genetics, e.g., for BRCA1, MLH1, and MSH2 analysis (11)(12), allows the detection of deletions or amplification of 50%. To our knowledge, however, analysis by MLPA of gradual loss of genetic material has not been published.
The identification of known tumor suppressor genes in specific tumors and at specific stages of carcinogenesis has prognostic value (13). Clinical diagnostics require methods for investigation of LOH of a tumor suppressor gene with high reproducibility, accuracy, and potential for automation. MALDI-TOF-MS-based genotyping of genetic heterogeneities (e.g., SNPs, insertions/deletions, microsatellite instabilities, and methylation) offers these features combined with excellent sensitivity [for a review see Ref. (14); also see Refs. (8)(15)(16)(17)(18)(19)]. The high sensitivity of MALDI-TOF-MS allows analysis of several genetic alterations in parallel.
Quantification of MALDI-TOF-MS signals requires low background and clear peak separation; therefore, use of deoxynucleotide triphosphates and dideoxynucleotide triphosphates for addition of one or two bases, respectively, producing mass differences of 300 Da, is a prerequisite for reproducible results. Only two enzymatic reactions are necessary in the described method, in contrast to chip-based technologies (20)(21)(22)(23), which require several modifications and/or hybridizations that may potentially alter the quantitative results. SNP-mapping array analysis detects the loss or addition of genetic material in cell lines (21)(22) but was verified just once in tumor tissue, and then without a discussion of homogeneity of tumor samples and gradual loss of genetic material (23).
In this report we show that semiquantitative MALDI-TOF-MS-based analysis of heterozygous SNPs is useful to confine (i.e., confirm alterations and investigate which of several genes in a region may be the disease-causing gene) genomic alterations at the gene level. This may lead directly to the identification of tumor suppressor genes or can be used as a prognostic and diagnostic tool to characterize known tumor suppressor genes.
Acknowledgments
We thank Dr. J. Brill for helpful discussions during manuscript preparation; S. Beck, B. Orlicz-Welcz, and P. Wenzeler for technical assistance; all of the patients who allowed the investigation; and the Department of Gynecology and Obstetrics, University Erlangen-Nürnberg (Prof. M. Beckmann) for providing the tissue samples. We are grateful to Dr. Markus Kostrzewa and Bruker Daltonik GmbH for helpful discussions and for providing the MALDI-TOF mass spectrometer. This work was supported by grants from Deutsche Forschungsgemeinschaft (DFG Schi 451/5-1), German-Israeli Foundation for Scientific Research & Development, and Marohn Stiftung, Erlangen.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and tumor tissue (
) DNA from patients 2, 20, 7, 10, and 22 were analyzed by MALDI-TOF-MS. Patients 2 and 20 had been classified by sequencing as showing LOH in tumor tissue, whereas patients 7, 10, and 22 had no LOH. Four to seven parallel extension reactions were analyzed. Mean (SD; error bars) signal ratios (T/C) are given for repeated analysis. The signal ratios for healthy and tumor tissue are significantly different in patients 2 and 20. In contrast, patients 7, 10, and 22 exhibit overlapping signal ratios for the healthy and tumor tissue. (C), signal ratios in healthy and tumor tissue from patient 2. Four heterozygous SNPs, exon 11 (C40678T; 
), exon 12 (C40907T; •), intron 11A (A40768G; 
), characterized by sequencing to exhibit LOH, were reanalyzed by MALDI-TOF-MS with up to four parallel extension reactions. Mean (SD; error bars) signal height ratios, if applicable, are given for repeated analyses. Statistical analysis for comparison of healthy and tumor tissue by t-test showed significant differences between signal ratios for tumor and healthy tissue (P <0.05). Signal ratios for an individual SNP were not significantly different within the healthy tissue group or within the tumor tissue group (P >0.05).