CGH and SNP array using DNA extracted from fixed cytogenetic preparations and long-term refrigerated bone marrow specimens
© MacKinnon et al; licensee BioMed Central Ltd. 2012
Received: 11 December 2011
Accepted: 2 February 2012
Published: 2 February 2012
The analysis of nucleic acids is limited by the availability of archival specimens and the quality and amount of the extracted material. Archived cytogenetic preparations are stored in many laboratories and are a potential source of total genomic DNA for array karyotyping and other applications. Array CGH using DNA from fixed cytogenetic preparations has been described, but it is not known whether it can be used for SNP arrays. Diagnostic bone marrow specimens taken during the assessment of hematological malignancies are also a potential source of DNA, but it is generally assumed that DNA must be extracted, or the specimen frozen, within a day or two of collection, to obtain DNA suitable for further analysis. We have assessed DNA extracted from these materials for both SNP array and array CGH.
We show that both SNP array and array CGH can be performed on genomic DNA extracted from cytogenetic specimens stored in Carnoy's fixative, and from bone marrow which has been stored unfrozen, at 4°C, for at least 36 days. We describe a procedure for extracting a usable concentration of total genomic DNA from cytogenetic suspensions of low cellularity.
The ability to use these archival specimens for DNA-based analysis increases the potential for retrospective genetic analysis of clinical specimens. Fixed cytogenetic preparations and long-term refrigerated bone marrow both provide DNA suitable for array karyotyping, and may be suitable for a wider range of analytical procedures.
KeywordsSNP array array CGH bone marrow archived specimens old archived specimens DNA extraction DNA analysis U937
Array comparative genomic hybridization (array CGH)  and single nucleotide polymorphism (SNP) array  are array-based karyotyping techniques which can help determine the genome abnormalities causing genetic disorders and the acquired genome copy number changes in cancer cells. They provide higher resolution than traditional karyotyping. However, the use of these and other DNA-based approaches for analysis is sometimes limited by the availability of suitable tissue samples, particularly for retrospective cancer genome analysis.
Fixed cytogenetic specimens are often stored after analysis, and are a potential source of total genomic DNA for array karyotyping and other DNA analysis techniques. Bone marrow specimens are commonly used to determine karyotype abnormalities in hematological malignancies. If unprocessed bone marrow is stored at 4°C during this time, it may be up to a month old before a karyotype is known and the decision to carry out array karyotyping is made. Anecdotally, it has been assumed that DNA extracted from these types of specimen is too degraded for analysis. Here we show that, on the contrary, an array karyotyping result can be obtained from these specimens.
The process of fixing and storing cells in 3:1 methanol/acetic acid introduces the possibility of acid nicking and degradation of the DNA. Two groups have reported array CGH using total genomic DNA extracted from cytogenetic preparations [3, 4]. To our knowledge this approach is not widely known.
Here we present a modified protocol for total genomic DNA extraction from fixed cytogenetic preparations, which addresses the need to obtain an optimum yield and concentration from a finite amount of starting material. We show that this DNA produces array CGH results of high quality, and we describe for the first time the use of DNA extracted from this source for SNP array. We also show that bone marrow that has been refrigerated (not frozen) for over a month yields DNA that, although partially degraded, produces reliable SNP array and array CGH results..
Selected copy number aberrations detected by SNP array and array CGH were validated by FISH (See Figures 1, 2, 3, 4 & 5). Chromosome 20 deletions were confirmed with a probe for the 20q12 deletion marker D20S108 (Vysis LSI D20S108 (20q12) SpectrumOrange). Chromosome 5q deletions were confirmed by multicolour FISH (M-FISH) and/or multicolour banding (M-BAND) and a probe detecting 5q31 deletions (Vysis LSI EGR1 (5q31) SpectrumOrange/D5S721, D5S23 SpectrumGreen, Abbott Molecular).
DNA size and yield
DNA yield from chromosome suspensions was in the range of 2-4 μg per 106 nuclei. Optical densities (OD260/280) were consistently at or above 1.8, which is the recommended purity for both the Agilent and Illumina microarray platforms. Qiagen recommends two 200 μL elutions for maximum yield from the DNeasy Cell and Tissue Kit. For low cellularity bone marrow and chromosome preparations, 200 μL elutions yielded DNA that was too dilute for the CGH or array protocols (e.g. < 10 ng/μL). We established that DNA could be extracted from a minimum of 106 fixed cells. By estimating the cellularity of chromosome preparations and using a lower first elution volume (40 μL) we were able to obtain DNA at a suitable concentration. From 1 × 106 fixed cells from a valuable specimen we obtained 69 ng/μL with a total yield greater than 4 μg.
Quality Control Metrics from DNA Workbench
Age of specimen1
Signal to Noise 3 Green
Signal to Noise 3 Red
DLRSpread values for array CGH from both fixed cytogenetic preparations (DLRSpread = 0.17-0.33) and longer-term refrigerated bone marrow specimens (DLRSpread = 0.16-0.28 for > 14 days at 4°C) compared favourably with those obtained using DNA extracted from fresher bone marrow (DLRSpread 0.14-0.20 for < 10 days at 4°C) (Table 1). The QC metrics for these specimens typically fell within the ranges "Good" to "Excellent", and included "Excellent" DLRSpread values for DNA extracted from bone marrow refrigerated for more than 26 days (Table 1). DNA extracted from bone marrow that had been refrigerated for 36 days also produced among the best DLRSpread values. The single specimen giving a "Poor" DLRSpread value obtained in this series of experiments, obtained from a one year old fixed cytogenetic preparation, also gave an "Excellent" DLRSpread value in a duplicate experiment; the overall result for this specimen was a "Pass". Therefore, the array CGH protocol can tolerate the level of DNA degradation that occurred in our specimens. Our results show that these specimens can be reliably used for array CGH.
There is no equivalent measure of SNP array signal spread or efficiency available in the Illumina Karyostudio software. However, SNP array images produced from both cytogenetic preparations and long-term refrigerated bone marrow showed low background and scatter (Figure 2).
Comparison of SNP array results from fresh and fixed cells
DNA was extracted from both live and fixed myeloid cell line U937 cells prepared from the same culture flask, for comparison of SNP array results. U937 contains copy number aberrations of all chromosomes except chromosome 9 from a basically triploid karyotype (; R. MacKinnon: A detailed molecular karyotype of the myeloid cell line U937 using combined FISH, M-FISH, M-BAND and SNP array, manuscript in preparation), allowing a comparison of the copy number aberrations and loss of heterozygosity (LOH) detected at many sites across the genome from fresh and fixed cells. The SNP array results displayed in Illumina Karyostudio matched the M-FISH pattern for this cell line (RM, unpublished results) and the copy number aberrations on chromosome 2 in U937 were confirmed by M-BAND (Figure 4).
Comparison of fresh and fixed specimen images for the same chromosomes revealed an increased spread of the B allele frequency (BAF) scatter plot from the fixed cells, which did not affect the result. It should be noted that the U937 cells were processed directly from live cells, whereas all bone marrow specimens analysed were stored at 4°C for at least one day before DNA extraction.
The same copy number aberrations were identified in fresh and fixed U937 tissue. There were minor variations in the boundaries called by the Karyostudio algorithm, an example of which is shown in Figure 4, in which a region of 2q gain was called as two distinct regions in the fresh specimen and a single region in the fixed specimen.
Some copy number aberrations could only be identified by visual examination of the B allele frequencies (an example shown in Figure 5). Mosaicism for a 7q deletion in the U937 specimen allowed us to assess the level of sensitivity of the SNP array for both fresh and fixed specimen. Two sub-clones in our U937 culture had overlapping deletions of 7q from one of four chromosomes 7. The overlapping deleted region (region a, Figure 5A) was identified in both the fresh and fixed U937 specimens, with BAF values of about 0.4 and 0.6 (equivalent to 66.7% cells with a 7q deletion on a tetrasomic background of AABB). The deletion was confirmed by FISH in metaphase nuclei and 207/300 (69%) interphase nuclei (Figure 5B). A deletion in 1/3 diploid cells would give the same BAF values. The larger deletion (region b, Figure 5A) was observed in 31% of metaphases (16/51), a frequency which would produce BAF values of 0.54 and 0.46 (equivalent to a deletion in 15% of diploid cells). This change in BAF values was apparent but not unequivocal in either specimen.
We have assessed the use of two types of archived specimens for both array CGH and SNP array. Cytogenetic preparations from diagnostic analysis are often archived in laboratory freezers and should be seen as a potential source of archival material for research into the genetics of malignancy and inherited disease. In our laboratory, fixed cytogenetic preparations are often the only stored patient specimen available. Bone marrow specimens received for karyotyping are also a potential source of DNA, but it has been assumed that the DNA is not suitable unless processed or frozen immediately. We have shown that these specimens can be successfully used for array karyotyping.
In 1986 Barker et al. reported that DNA suitable for Southern analysis can be extracted from fixed cytogenetic preparations, using a phenol/chloroform protocol, and they suggested that this DNA might be suitable for other analytical purposes. Our array CGH results confirm the use of DNA extracted from fixed cytogenetic preparations for array CGH [3, 4, 7], and here we show that the quality metrics of the results compare favourably with results from one day old bone marrow. We have shown for the first time that DNA from fixed cytogenetic preparations can also be used for SNP array. DNA prepared from fixed chromosomes was slightly degraded but was not degraded further with longer storage at -80°C (Figure 6).
We describe a modified DNA extraction method for use with cytogenetic suspensions. By processing a known cell number and performing an initial lower volume elution, the amount of specimen used can be kept to a minimum. This is important for limited volume specimens or specimens with low cellularity. In low cellularity specimens this produced an eluant that could be used without further concentration and loss of DNA. Interestingly we have managed to extract a small amount of highly degraded RNA from fixed cytogenetic suspensions which have been used successfully for Real-Time PCR .
Validation of array CGH and SNP array results showed that they were reliable. Direct comparison of the same specimen processed fresh or fixed showed only a small increase in the spread of B allele frequencies in results derived from a fixed specimen when compared with the same cell line processed fresh. A deletion producing a shift in BAF values from 0.5 to 0.4 and 0.6 was clearly visible in both specimens (equivalent to a deletion in 1/3 of diploid cells).
We also achieved reliable SNP array and array CGH results using unprocessed bone marrow specimens refrigerated for 36 days or more before freezing. Agarose gel analysis showed that the DNA in bone marrow stored at 4°C degraded over time (see Figure 6). However, specimens stored for up to 36 or 42 days at 4°C were still suitable for both array CGH and SNP array, respectively.
As it may occasionally take weeks to determine a karyotype, considerable time and resources can be saved if only the specimens of interest are frozen or processed after a karyotype is known. This approach is particularly suitable for bone marrow, which is difficult to re-collect. Also, therapy may have been administered or the genotype of malignant cells may have changed by the time a subsequent collection is contemplated.
Cytogenetic slides are another potential source of DNA for microarray studies. A method has been described for extracting DNA from chromosomes scraped from microscope slides for PCR . However, around 106 nuclei are needed for the extraction protocol we describe, and so a single cytogenetic slide would not produce enough DNA for array karyotyping without amplification. Thus, while DNA could potentially be obtained from slides, use of unspread specimens is simpler, and is less likely to require amplification.
Other methods of DNA analysis may also be possible using DNA extracted from these specimens. DNA extracted from formalin-fixed paraffin embedded tissues (FFPE) can be used for array CGH and SNP array analysis, although sensitivity is much poorer than for fresh tissue [10–12]. Whole genome amplification makes it possible to use a small starting amount of DNA [10, 12]. Improved protocols make sequencing of degraded DNA extracted from FFPE specimens possible [13, 14], suggesting that DNA from cytogenetic preparations will be even more suitable for this and other methods of DNA analysis. Massively parallel sequencing can be carried out on as few as six microdissected and amplified chromosome segments .
We have shown that cytogenetic preparations in long-term storage, and bone marrow specimens which have been refrigerated unprocessed for at least 36 days, can be used as a source of genomic DNA for both SNP array and array CGH. We have also described a modified DNA extraction protocol for use with cytogenetic preparations of low cellularity. These methods will prove particularly useful for cancer genome analysis. They will allow chromosome abnormalities at a certain point in disease evolution to be studied retrospectively, and make retrospective analysis possible for patients for whom there are no alternative archival specimens.
All DNA extractions were performed with a DNeasy Cell and Tissue Kit (Qiagen, Germantown, MD), using the blood protocol or a modification thereof.
Bone marrow from patients with myeloid malignancies was collected in tubes containing 100 IU (1,000 IU/mL) sodium heparin and sent to the Victorian Cancer Cytogenetics Service (VCCS) for cytogenetic analysis. Residual specimen that was not used for the preparation of fixed cell suspensions was stored at 4°C. After determination of the karyotype, residual bone marrow specimens of potential use for array karyotyping studies were transferred to cryotubes and stored at -80°C. Total genomic DNA was prepared subsequently from 100 μL thawed whole bone marrow specimen.
Total genomic DNA was also prepared from fixed cytogenetic suspensions which had been prepared from cultured bone marrow cells using standard techniques (hypotonic treatment followed by lysis, the addition of 3:1 methanol:acetic acid ) and stored at -80°C for up to twelve years. Cell concentration was estimated by thorough resuspension and spreading of three 1 μL aliquots from separate known dilutions on a clean glass slide. Interphase and metaphase nuclei were counted and averaged. Cytogenetic suspensions containing at least 106 nuclei were rinsed three times with PBS, resuspended in 200 μL PBS and processed using the DNeasy Cell and Tissue Kit without RNase treatment. However, when specimens were of low cellularity, the recommended 200 μL elution produced specimens which were too dilute for accurate quantitation and direct use in the array protocols. Therefore, instead of the recommended 200 μL elution, two or more elutions of 40-100 μL, for at least 10 minutes each, were carried out with the AE eluant provided in the kit. The lower volume first elution was chosen if the amount of specimen was limited (less than 107 nuclei), to ensure a usable final concentration. Further elutions were used to recover more of the residual DNA from the column.
The U937 cell line [5, 16] (obtained from the laboratory of Hamish Scott, Walter and Eliza Hall Institute, Melbourne), was cultured in RPMI containing 10% FCS, glutamine, penicillin and streptomycin at 37°C in air containing 5% CO2. Fresh cultured U937 cells were split into two equal volumes and immediately processed. DNA was extracted by two different methods, to allow a comparison between DNA extracted from fresh and fixed tissue: (1) DNA was extracted directly from cultured cells using the Qiagen DNeasy kit blood protocol, or (2) metaphases were harvested according to standard cytogenetics protocols, stored at -80°C for 70 days, and DNA extracted with the DNeasy Cell and Tissue kit using the protocol described above.
DNA was run on 0.7% agarose to check for integrity and quantitated with a Nanodrop spectrophotometer (Thermo Scientific, Wilmington, DE).
Array karyotyping - SNP array
Illumina CytoSNP 12 arrays (Illumina, San Diego, CA) were processed according to the manufacturer's instructions using 200 ng of each DNA sample in a 4 μL volume. Data were analysed using Karyostudio (version 1.2, Illumina).
B allele frequencies were used to determine copy numbers of chromosome 2 regions in the abnormal chromosomes represented in Figure 4. The SNP array pattern (log R ratio and B allele frequency) is consistent with (from p arm to q arm) 3-2-3-4-2 copies of each region. The B allele frequencies show that at each region there is one copy of one homolog and 1-3 copies of the other homolog, which is consistent with all rearrangements occurring in one of the homologs (Figure 4).
Array karyotyping - array CGH
Array CGH was carried out using the Agilent platform according to the manufacturer's instructions. Custom 44K and 105K arrays with a high probe density on chromosome 20 were used for all but the 27 day old bone marrow specimen (see Table 1) which was a catalog 60K array (design 021924). Test and control (Promega pooled Human Genomic DNA of the opposite sex) DNA were labelled with Cyanine 3-dUTP and Cyanine 5-dUTP using an Agilent Genomic DNA Labeling Kit PLUS (Agilent, Santa Clara, CA), denatured and co-hybridized before hybridizing to a custom 44K array. Data were analysed using the Genomic Workbench software (version 5.0.14, Agilent) and regions of significant gain or loss were determined using a z-score algorithm with a threshold of 2.5. Most experiments were performed in duplicate by swapping dyes between test and control.
Validation of array results
Copy number aberrations identified by array CGH and SNP array were validated by FISH, M-FISH and M-BAND using protocols described in MacKinnon et al.. Metasystems XCyte probes (Metasystems, Altlusshem, Germany) for M-FISH and M-BAND were used according to the manufacturer's instructions.
M-BAND of chromosome 2 was used to confirm the copy number aberrations detected in the cell line U937, using the XCyte 2 probe (Metasystems, Altlussheim, Germany). M-BAND probes are chromosome-specific and produce a multicolor banded chromosome pattern. The XCyte 2 probe is comprised of eight partially overlapping region-specific paints labeled with different fluorochromes. The Isis algorithm converts relative fluorescence intensities into false colors, creating a multicolored banding pattern where each region of chromosome 2 is identified by a unique color. Comparison of the pattern on the abnormal chromosomes with the normal chromosome 2 pattern showed deletion of 2qter and inverted duplication of the adjacent segment in one abnormal chromosome and partial loss of 2p in the other.
Deletions of 20q were confirmed by FISH with a probe for the common deleted region (Vysis LSI D20S108 (20q12) SpectrumOrange, Abbott Molecular, Abbott Park, Ill.). Chromosome 5 deletions were confirmed by M-FISH and/or M-BAND and a Vysis probe detecting 5q31 deletions (LSI EGR1 (5q31) SpectrumOrange/D5S721, D5S23 SpectrumGreen, Abbott Molecular). Deletion of 7q was confirmed by Vysis LSI D7S486 (7q31) SpectrumOrange/CEP7 SpectrumGreen (Abbott Molecular).
This study was approved by Human Research Ethics Committee A of St Vincent's Hospital (Melbourne) Ltd, Protocol HREC-A 091/02, and complies with the Helsinki Declaration.
Grants: Funded by ANZ Medical Research and Technology in Victoria, the James and Vera Lawson Philanthropic Trust (ANZ Trustees), the Leukaemia Foundation of Australia and the Cancer Council of Victoria. We thank Ella Wilkins and Hamish Scott for use of the cell line U937, Dan Belluoccio of Agilent Technologies for processing the 60k Agilent array, and Derek Campbell of Illumina for assistance with the SNP arrays.
- P Pinkel D, Segraves R, Sudar D, Clark S, Poole I, Kowbel D, Collins C, Kuo WL, Chen C, Zhai Y, Dairkee SH, Ljung BM, Gray JW, Albertson DG: High resolution analysis of DNA copy number variation using comparative genomic hybridization to microarrays. Nat Genet 1998, 20: 207–11. 10.1038/2524View ArticlePubMedGoogle Scholar
- Mei R, Galipeau PC, Prass C, Berno A, Ghandour G, Patil N, Wolff RK, Chee MS, Reid BJ, Lockhart DJ: Genome-wide detection of allelic imbalance using human SNPs and high-density DNA arrays. Genome Res 2000, 10: 1126–37. 10.1101/gr.10.8.1126PubMed CentralView ArticlePubMedGoogle Scholar
- Evers C, Beier M, Poelitz A, Hildebrandt B, Servan K, Drechsler M, Germing U, Royer HD, Royer-Pokora B: Molecular definition of chromosome arm 5q deletion end points and detection of hidden aberrations in patients with myelodysplastic syndromes and isolated del(5q) using oligonucleotide array CGH. Genes Chromosomes Cancer 2007, 46: 1119–28. 10.1002/gcc.20498View ArticlePubMedGoogle Scholar
- Yu S, Bittel DC, Kibiryeva N, Zwick DL, Cooley LD: Validation of the Agilent 244K oligonucleotide array-based comparative genomic hybridization platform for clinical cytogenetic diagnosis. Am J Clin Pathol 2009, 132: 349–60. 10.1309/AJCP1BOUTWF6ERYSView ArticlePubMedGoogle Scholar
- Shipley JM, Sheppard DM, Sheer D: Karyotypic analysis of the human monoblastic cell line U937. Cancer Genet Cytogenet 1988, 30: 277–84. 10.1016/0165-4608(88)90195-1View ArticlePubMedGoogle Scholar
- Barker PE, Testa JR, Parsa NZ, Snyder R: High molecular weight DNA from fixed cytogenetic preparations. Am J Hum Genet 1986, 39: 661–8.PubMed CentralPubMedGoogle Scholar
- MacKinnon RN, Kannourakis G, Wall M, Campbell LJ: A cryptic deletion in 5q31.2 provides further evidence for a minimally deleted region in myelodysplastic syndromes. Cancer Genet 2011, 204: 187–94. 10.1016/j.cancergen.2011.02.001View ArticlePubMedGoogle Scholar
- MacKinnon RN, Selan C, Wall M, Baker EG, Nandurka H, Campbell LJ: The Paradox of 20q11.21 Amplification in a Subset of Cases of Myeloid Malignancy with Chromosome 20 Deletion. Genes Chromosomes Cancer 2010, 48: 998–1013.View ArticleGoogle Scholar
- Jonveaux P: PCR amplification of specific DNA sequences from routinely fixed chromosomal spreads. Nucleic Acids Res 1991, 19: 1946. 10.1093/nar/19.8.1946PubMed CentralView ArticlePubMedGoogle Scholar
- Little SE, Vuononvirta R, Reis-Filho JS, Natrajan R, Iravani M, Fenwick K, Mackay A, Ashworth A, Pritchard-Jones K, Jones C: Array CGH using whole genome amplification of fresh-frozen and formalin-fixed, paraffin-embedded tumor DNA. Genomics 2006, 87: 298–306. 10.1016/j.ygeno.2005.09.019View ArticlePubMedGoogle Scholar
- Thompson ER, Herbert SC, Forrest SM, Campbell IG: Whole genome SNP arrays using DNA derived from formalin-fixed, paraffin-embedded ovarian tumor tissue. Hum Mutat 2005, 26: 384–9. 10.1002/humu.20220View ArticlePubMedGoogle Scholar
- Tuefferd M, De Bondt A, Van Den Wyngaert I, Talloen W, Verbeke T, Carvalho B, Clevert DA, Alifano M, Raghavan N, Amaratunga D, Gohlmann H, Broet P, Camilleri-Broet S: Genome-wide copy number alterations detection in fresh frozen and matched FFPE samples using SNP 6.0 arrays. Genes Chromosomes Cancer 2008, 47: 957–64. 10.1002/gcc.20599View ArticlePubMedGoogle Scholar
- Wood HM, Belvedere O, Conway C, Daly C, Chalkley R, Bickerdike M, McKinley C, Egan P, Ross L, Hayward B, Morgan J, Davidson L, MacLennan K, Ong TK, Papagiannopoulos K, Cook I, Adams DJ, Taylor GR, Rabbitts P: Using next-generation sequencing for high resolution multiplex analysis of copy number variation from nanogram quantities of DNA from formalin-fixed paraffin-embedded specimens. Nucleic Acids Res 2010, 38: e151. 10.1093/nar/gkq510PubMed CentralView ArticlePubMedGoogle Scholar
- Schweiger MR, Kerick M, Timmermann B, Albrecht MW, Borodina T, Parkhomchuk D, Zatloukal K, Lehrach H: Genome-wide massively parallel sequencing of formaldehyde fixed-paraffin embedded (FFPE) tumor tissues for copy-number- and mutation-analysis. PLoS ONE 2009, 4: e5548. 10.1371/journal.pone.0005548PubMed CentralView ArticlePubMedGoogle Scholar
- Weise A, Timmermann B, Grabherr M, Werber M, Heyn P, Kosyakova N, Liehr T, Neitzel H, Konrat K, Bommer C, Dietrich C, Rajab A, Reinhardt R, Mundlos S, Lindner TH, Hoffmann K: High-throughput sequencing of microdissected chromosomal regions. Eur J Hum Genet 2010, 18: 457–62. 10.1038/ejhg.2009.196PubMed CentralView ArticlePubMedGoogle Scholar
- Ralph P, Harris PE, Punjabi CJ, Welte K, Litcofsky PB, Ho MK, Rubin BY, Moore MA, Springer TA: Lymphokine inducing "terminal differentiation" of the human monoblast leukemia line U937: a role for gamma interferon. Blood 1983, 62: 1169–75.PubMedGoogle Scholar
- Moorhead PS, Nowell PC, Mellman WJ, Battips DM, Hungerford DA: Chromosome preparations of leukocytes cultured from human peripheral blood. Exp Cell Res 1960, 20: 613–6. 10.1016/0014-4827(60)90138-5View ArticlePubMedGoogle Scholar
- MacKinnon RN, Patsouris C, Chudoba I, Campbell LJ: A FISH comparison of variant derivatives of the recurrent dic(17;20) of myelodysplastic syndromes and acute myeloid leukemia: Obligatory retention of genes on 17p and 20q may explain the formation of dicentric chromosomes. Genes Chromosomes Cancer 2007, 46: 27–36. 10.1002/gcc.20385View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.