Introduction
The P53 tumor suppressor gene is one of the most commonly mutated genes in human cancers (1). The majority of P53 mutations are missense mutations, generally located in the central domain of the P53 coding region (exons 5–8) (2,3). Other tumor suppressor genes predominantly contain nonsense or other truncation-producing alterations (4–6).
P53 mutations have been identified in a wide variety of human cancers, including invasive breast duct carcinoma (IDC), acute myeloblastic leukaemia (AML), astrocytoma and soft tissue sarcoma (STS). P53 mutations have been identified in 15–50% of IDC (7), in 4.5–9% of AML (8,9), in 12.5–59% of astrocytoma (10,11) and in 11–20% of STS (12,13). The wide range of P53 mutation frequencies in these cancers may be due to cancer tissue heterogeneity, differences in the examined population or the technical approach employed. The presence of P53 mutations may have both prognostic and therapeutic consequences. Therefore, the use of the optimal method for the determination of P53 mutations is critical. The P53 mutation has been reported to be of prognostic importance in breast cancer, AML, astrocytoma and STS (8,14–17). Thus, detection of the actual incidence of P53 mutations is essential in these cancers.
Methods used for the detection of P53 mutations are based either on genomic DNA or mRNA as a template (11,12,15). The most widely used methods are based on DNA sequencing. However, few studies exist that compare sequencing assays by using both RNA and DNA targets (18–22). Recently, it was demonstrated that the detection of P53 mutations has improved, and P53 mutations are detected more frequently at the mRNA level (using cDNA sequencing) than at the DNA level in glioblastomas, colorectal cancer and pleomorphic xanthoastrocytoma (20–22). Therefore, whether similar discrepancies occur in other types of human cancers was investigated. For this study, IDC, AML, astrocytoma and STS were selected as the cancers presenting frequent P53 mutations. The P53 gene status was evaluated by means of both cDNA and DNA sequencing and real-time quantitative RT-PCR.
Materials and methods
Tumor samples
The study included 40 patients with IDC, 23 patients with AML, 12 patients with astrocytoma (7 cases of anaplastic astrocytoma WHO grade III, 5 cases of diffuse astrocytoma WHO grade II) and 40 patients with STS. The patients with STS, IDC and astrocytoma were diagnosed at the Department of Pathology, Medical University of Lodz. The patients with AML were diagnosed at the Department of Haematology, Medical University of Lodz. Approval for the study was provided by the Bioethical Committee of Medical University of Lodz (No. RNN/53/08/KE). Patient consent was obtained from each individual included in the study.
DNA and RNA isolation
Isolation was performed using snap-frozen tissues stored at −80˚C and fresh or frozen leucocytes of peripheral blood or bone marrow cells obtained from the patients, prior to the initiation of the therapy. DNA and RNA were co-extracted by means of the Macherey-Nagel DNA/RNA purification kit. RNA samples were treated with DNase. RNA and DNA concentrations were measured spectrophotometrically. Total RNA (100 ng) was reverse-transcribed into single-stranded cDNA in a final volume of 40 μl containing 50 mM DTT, 1.5 μg oligo(dT), 0.5 mM dNTP, 40 units RNase inhibitor and 200 units M-MLV reverse transcriptase (Promega).
P53 DNA and cDNA sequencing
Exons 5–8 of the P53 gene were amplified by PCR as previously described and sequenced using the dideoxy termination method and the SequiTherm Excel DNA Sequencing kit (Epicentre Technologies) (20,23). The primer sequences are shown in Table I. DNA and cDNA sequence alterations were reconfirmed by a repeated analysis of isolated DNA and RNA, respectively. To verify the results of sequencing, the semi-quantitative densitometric analysis was performed as previously described (21). The intensity of wild-type and mutated bands was estimated by comparison with neighbouring bands in the same sequencing lane, and used as a reference (21).
Real-time quantitative RT-PCR
Real-time quantitative RT-PCR was performed on a Rotor Gene 6000 instrument (Corbett, Life Sciences, Australia) for the P53 gene (TaqMan® Gene Expression Assays no. Hs00153340_m1 and no. Hs00153349_m1), with GAPDH (TaqMan® Gene Expression Assays no. Hs99999905_m1) as a reference gene for the normalization of the target gene expression levels. The normalized relative P53 expression level in an unknown sample vs. the control sample was calculated using the method described by Pfaffl et al with the blood of two healthy donors as control tissue (24).
Statistical analysis
The differences in P53 expression levels for a particular cancer type and for all cancer types together were evaluated by the Mann-Whitney U test. Statistical significance was set at P≤0.05. The statistical analysis was performed with Stat-View for Windows 5.01 Software® (SAS Institute Inc., Cary, NC, USA).
Results
cDNA sequencing in IDC discerns the P53 mutation more frequently than DNA sequencing
Sequencing of DNA showed P53 mutations in 3 of the 40 analyzed IDC cases (7.5%). Sequencing of cDNA showed the presence of a P53 mutation in 5 more cases, i.e., 8 of 40 (20%). One deletion and 4 missense mutations were found on the cDNA sequencing, but not on the DNA sequencing (cases nos. 1, 2, 6, 7 and 8). Three cases showed a P53 mutation on both the DNA and cDNA sequence: 1 missense mutation with a visible wild-type nucleotide on the DNA sequence showed only the mutated cDNA template with the elimination of wild-type mRNA (case no. 3); 1 missense mutation with visible wild-type nucleotide on the DNA sequence showed a higher ratio of mutated to wild-type on the cDNA template (case no. 4) and 1 case did not differ in the presence of a mutation between the DNA and cDNA sequence (case no. 5; Table II). The detection of P53 mutations was found to be 2.7 times more frequent in IDC on the cDNA vs. the DNA sequence.
cDNA sequencing in AML discerns the P53 mutation more frequently than DNA sequencing
The P53 DNA sequencing results for AML, in part, were previously reported (25). Sequencing of DNA showed the presence of a P53 mutation in 1 of the 23 AML cases (4.3%). Sequencing of cDNA showed a P53 mutation in 1 more case, i.e., 2 of 23 (8.7%). A missense mutation based only on cDNA was noted in case no. 2; in case no. 3 no difference in the presence of mutations between the DNA and cDNA sequence was observed and case 1 showed the A/G polymorphism at codon 213 on both the cDNA and DNA sequences (Table II). The detection of the P53 mutation was found twice as frequently in AML using cDNA vs. DNA sequencing.
cDNA sequencing in astrocytoma discerns the P53 mutation more frequently than DNA sequencing
Sequencing of DNA showed the presence of P53 mutations in 2 of the 12 astocytoma cases (16.7%). Sequencing of cDNA showed P53 mutations in 1 more case, i.e., 3 of 12 (25%). One mutation (deletion) was found based only on cDNA sequencing (case no. 3), and 2 mutations did not differ between the DNA and cDNA sequence (Table II). The detection of P53 mutations was 1.5 times more frequent in astrocytoma using cDNA vs. DNA sequencing.
cDNA sequencing in STS discerns the P53 mutation more frequently than DNA sequencing
The P53 DNA sequencing results in STS, in part, were previously reported (26). Sequencing of DNA showed the presence of P53 mutations in 8 of the 40 STS cases (20%). Sequencing of cDNA showed a P53 mutation in 1 more case, i.e., 9 of 40 (22.5%). One missense mutation was found based only on cDNA sequencing (case no. 8), and 2 missense mutations with a visible wild-type nucleotide on DNA sequence showed only the mutated cDNA template with the elimination of wild-type mRNA (cases no. 9 and 10; Table II). Six missense mutations did not differ between the DNA and cDNA sequences. Six polymorphism sites were found on both the cDNA and DNA sequences (Table II). The detection of P53 mutations was 1.1 times more frequent in STS using cDNA sequencing than using DNA sequencing.
cDNA sequencing in all of the cancer types discerns the P53 mutation more frequently than DNA sequencing
A total of 14 mutations were found in both the cDNA and genomic DNA sequences. Eight additional mutations (6 missense and 2 deletions) were found based only on cDNA sequencing (Table II). No mutations were found only on the DNA sequence. P53 mutations were detected more frequently in the examined cancer types when the cDNA sequencing assay was used, i.e., 8.3% and 5.3% of P53 mutations were found based on the cDNA and DNA sequence, respectively. The highest difference in the frequency of P53 mutation detection using cDNA and DNA sequencing assay was observed in IDC. Furthermore, 4 of the 10 samples presented a missense P53 mutation with visible wild-type nucleotide on the DNA sequence which showed elimination of the wild-type allele (3 samples) or selective overproduction of the mutated allele on the cDNA sequence (1 sample) (Table II).
P53 expression did not differ between cases with and without P53 mutations
The mean P53 mRNA level did not increase in any of the cancers examined in comparison to the control blood samples (the mean level for each type of cancer was ~1). P53 mRNA levels were compared between the two groups: cases with P53 mutations and those lacking a P53 mutation. P53 expression was not significantly different between the two groups (P>0.05 for a particular cancer type and for all cancer types together).
P53 expression was ~2-fold higher, but not significantly, in samples that presented a missense P53 mutation with a visible wild-type nucleotide on DNA and with the elimination of the wild-type allele or selective overproduction of the mutated allele on the cDNA sequence (mean fold 1.9) than in samples which showed a similar level of mutated and wild-type template on both the DNA and cDNA sequence (mean fold 0.9).
Discussion
This is a comparative study of two sequencing assays, cDNA and DNA sequencing, in IDC, AML, astrocytoma and STS. An analysis of these cancers supports our previous data for glioblastoma, colorectal cancer and pleomorphic xanthoastrocytoma. Subsequently, the detection of P53 mutations is improved and more frequent at the mRNA than at the DNA level (20–22). Our previous comparative studies showed that the selective promotion of the mutated P53 mRNA in glioblastoma samples with P53 heterozygous missense mutations, and P53 missense mutations in colorectal cancer and pleomorphic xanthoastrocytoma were more frequently detected using cDNA vs. DNA sequencing (20–22). A comparative study of P53 cDNA and DNA sequencing in colorectal cancer by Forslund et al showed a higher, although not significant, frequency of P53 mutation detection using cDNA sequencing as opposed to DNA sequencing (10 mutations were found based only on the cDNA sequence and 3 mutations were found based only on the DNA sequence) (19).
In the present study, the P53 mutation was detected more frequently in all of the cancer types when the cDNA sequencing assay was used, whereas a total of 8 additional mutations (6 missense and 2 deletions) were found based only on the cDNA sequence. No mutations were detected on the DNA sequence alone. The highest difference in the frequency of P53 mutation detection using the cDNA and DNA sequencing assay was observed in IDC. Moreover, several samples presented a missense P53 mutation with a visible wild-type allele on DNA sequencing, that may have originated from infiltrating or surrounding normal cells, or may have been a heterozygous missense mutation. In contrast, elimination of the wild-type allele or selective overproduction of the mutated allele was noted on the cDNA sequence. A similar phenomenon was observed in our previous study of glioblastoma with heterozygous missense mutations, in which elimination of the wild-type allele was observed. In addition, a study of breast cancer with missense mutations also presented the wild-type DNA nucleotide, which originated from infiltrating or surrounding normal cells in which a higher tumor-specific signal for RNA than for DNA was detected (18,20). The reason for the discrepancies between DNA and cDNA P53 analysis may be, at least in part, the genotypical heterogeneity of cells located in cancer tissues (21,22). The contamination of tumor samples by cells without P53 mutations (normal and/or neoplastic) results in difficulties in the detection of P53 mutations using DNA analysis. In a sample containing less than 25–30% of tumor cells, P53 mutation detection using direct DNA sequencing is almost impossible (21,27). However, for cases presenting heterozygous missense mutations, the mechanisms responsible for the selective overexpression of mutated P53 mRNA require further scrutiny. The investigation of glioblastoma with heterozygous missense mutations enabled us to propose three possible hypotheses to explain these discrepancies: i) the silencing of wild-type mRNA transcription; ii) the degradation of wild-type mRNA; and iii) the selective overproduction of mutated mRNA (21). The mechanism(s) which favour the mutated type of P53 mRNA have yet to be determined. Nevertheless, our approach, based on cDNA analysis, increased P53 mutation detectability for the examined cancers. Moreover, we cannot exclude the hypothesis that the up-regulation of the mutated P53 allele co-exists with the silencing of the wild-type allele in these cancers. The use of an optimal method for the determination of P53 mutations is important, particularly for cancers in which the presence of P53 mutations is a diagnostic, predictive and prognostic factor. It should be kept in mind that, although we employed the cDNA sequencing assay in this study, it is very sensitive in the detection of the mutated template since it does not allow for the detection of all types of mutations (18,28).
In colorectal cancer an increased mRNA P53 expression and no difference in the mRNA P53 expression between cancers with or without a P53 mutation has been reported (21). In our study, the P53 mRNA level did not increase in any type of cancer examined and was not significantly different between cases with or without a P53 mutation.
In conclusion, we suggest that the true incidence of P53 mutations in the examined cancers is underestimated at the DNA level, and evaluation of the gene alteration should be conducted using cDNA analysis. The approach adopted by the authors was based on cDNA analysis and resulted in an increased detection of P53 mutations in all types of analyzed cancers. Thus, we suggest that P53 cDNA sequencing be applied for these cancers to increase P53 mutation detectability.