Assessment of sequence-based p53 gene ... - Semantic Scholar

Clinical Chemistry 44:3 455– 462 (1998)

Molecular Pathology and Genetics

Assessment of sequence-based p53 gene analysis in human breast cancer: messenger RNA in comparison with genomic DNA targets Cecilia Williams,1 Torbjo¨rn Norberg,2 Afshin Ahmadian,1 Fredrik Ponteń,3 Jonas Bergh,4 Mats Ingana¨s,2 Joakim Lundeberg,1* and Mathias Uhleń1

uncontrolled expansion of the tumor, because generally a mutated p53 protein can neither attain its function as a check-point controller nor induce apoptosis or cellular repair. Loss of heterozygosity (LOH) in the 17p13 region harboring the p53 gene is observed in ;60% of breast cancer tumors, and only one-half or fewer of these show the second allele to be mutated [3–5]. Overexpression of p53 protein is another property frequently observed in breast cancer cells where the p53 gene is mutated, and is associated with poor prognosis [6 – 8]. The common explanation for this phenomenon is a prolonged half time of elimination for mutated proteins [9]. However, there are several conflicting examples [10, 11]. For instance, overexpressed native p53 protein appears in many tumors [3] as well as in normal cells after exposure to DNA-damaging agents [12]. In addition, Li–Fraumeni patients with one mutated p53 allele do not exhibit increased p53 protein concentrations in normal cells, but the same mutant accumulates to high amounts in tumor cells of the same patient [13]. The high prevalence of p53 mutations in human cancers and the suggestion that the presence or absence of p53 mutations might have both prognostic and therapeutic consequences [14 –16] make direct sequencing methods for this gene attractive. Strategies have been described based on either mRNA [15] or genomic DNA [17] as a template. In PCR–single-strand conformation polymorphism, the number of mutations detected has been shown to be influenced by whether RNA or DNA was used as a target [18]. Despite this, no comparative study between these two sequencing methods has been performed. In this report, we compare and evaluate sequencing assays by using RNA and DNA targets for various types of mutations and correlate their relationships with other variables, such as immunohistochemistry (IHC), LOH, and steady-state concentrations of mutated p53 mRNA.

The high prevalence of p53 mutations in human cancers and the suggestion from several groups that the presence or absence of p53 mutations might have both prognostic and therapeutic consequences point to the importance of optimal methods for p53 determination. Several strategies exploring this have been described, based either on mRNA or genomic DNA as a template. However, no comparative study on the reliability of the two templates has been performed. The principal aim of this study was to study the concordance of RNA- and DNA-based direct sequencing methods in detecting p53 mutations in breast tumors. In 100 tumors, 22 mutations were detected by both methods. Furthermore, one stop mutation, two splice-site mutations, and one intron alteration were found only by genomic sequencing. In addition, the comparative study suggests that cells with missense mutations have increased steady-state concentrations of p53-specific mRNA, in contrast to cells with a gene encoding a truncated protein. Mutations in the p53 tumor suppressor gene are found at high frequency in a wide range of human cancers. In breast cancer, p53 mutations appear to be an early event in the progression of cancer [1] and occur in ;22% [2] of malignant breast tumors. The majority of these mutations occur as point mutations in the evolutionary conserved regions of the p53 gene, affecting the DNAbinding area of the protein. Mutation of one allele combined with loss of the second allele may result in

1 Department of Biochemistry and Biotechnology, Royal Institute of Technology (KTH), S-100 44 Stockholm, Sweden. 2 Pharmacia Biotech AB, S-751 82 Uppsala, Sweden. Departments of 3 Pathology and 4 Oncology, Akademiska University Hospital, S-751 85 Uppsala, Sweden. * Author for correspondence. Fax 46-8-24 54 52; e-mail joakim.lundeberg@ biochem.kth.se. Received April 28, 1997; revision accepted October 10, 1997.

455

456

Williams et al.: Sequence-based p53 gene analysis

Materials and Methods Tissue specimens. One hundred breast cancer tumors were selected for genomic p53 analysis from a previous study of 316 patients relating prognostic information to p53 mRNA analysis [19]. After rejection of the smallest tumors, the 100 samples were selected at the Department of Pathology (University Hospital, Uppsala, Sweden). The IHC, S-phase, and aneuploidy data related to these samples have been published previously [15]. In short, for IHC staining, the anti-p53 monoclonal antibody 1801 (Biozac AB), which identifies both the wild-type and mutant p53 protein, was used, and the degree of staining was estimated according to the extent and intensity of staining by using a scale ranging from 0 (negative) to 9 (strong overall staining). A sample was considered positive if it stained as grades 1–9, irrespective of intensity or frequency of staining. The S-phase fraction was classified as high or low (cutoff point: 7% for diploid and 12% for aneuploid tumors, respectively). Sample preparation. The method is outlined in Fig. 1. To allow for analysis of tumor genomic DNA, microdissection was used to minimize the amount of stroma cells surrounding the infiltrating tumor cells (as shown in Fig. 2). Microdissection and sample preparation were performed as described previously [20]. Briefly, tumor nests containing as few as 50 cells were isolated from 16-mm cryostat sections stained with methylene blue. After microdissection, where normal cells were removed, the cells to be analyzed were transferred to tubes containing 50 mL of 10 mmol/L Tris-HCl, pH 8.3, and 50 mmol/L KCl. The microdissected cells were lysed by the addition of 2 mL of proteinase K (20 mg/mL) at 56 °C for 1 h. Proteinase K was heat-inactivated (95 °C for 10 min) before outer PCR was performed. Preparation of tumor for RNA sample analysis (Fig. 1) was performed by a total RNA extraction procedure, without microdissection, as described previously [15]. In short, a thin slice (;5 3 2 3 2 mm) of the frozen tumor was cut and transferred to extraction solution (RNAzole®; Cinna Biotec, Inc.). The cells were disrupted by grinding with a micro pestle. The extraction was performed according to the method of Chevillard [21] and Chomczynski and Sacchi [22] and finally the precipitated RNA was dissolved in 50 mL of diethyl pyrocarbonate-treated water containing 25 units of RNA Guard® (Pharmacia Biotech AB). To prepare 25 mL of cDNA for the RNA, samples were heated to 90 °C for 3 min and transferred to ice. Moloney murine leukemia virus reverse transcriptase, RNA Guard, and cDNA-mix were subsequently added according to the method of Sjo¨gren et al. [15]. The reaction mix was incubated at 37 °C for 1 h and finally heat-denatured at 90 °C for 3 min and stored at 220 °C until PCR reactions were performed. Thus, no precautions were taken to avoid normal cells in the mRNA analysis. The samples for DNA analysis contained ;50 –500 cells, whereas the RNA samples contained at least 1000-fold more cells.

Sample preparation of control samples. In a set of 16 control samples, genomic DNA and mRNA targets were processed in parallel by coextraction directly from the biopsy samples (without initial microdissection), resulting in exactly the same cell population being analyzed in the two assays. Five tumors were randomly selected from the 22 tumors containing mutations, as determined by both assays, and 11 additional, unrelated breast cancer samples were included that had been identified previously as mutated by the RNA target assay [19]. By coextraction according to the method of Chevillard [21], DNA and RNA were separated into different phases and then analyzed in the same manner as the previous samples (starting from multiplex PCR for DNA and cDNA synthesis for RNA). Thus, a direct reflection of genomic DNA and corresponding mRNA content was achieved. LOH analysis. Chromosomal DNA was analyzed for LOH by using polymorphic markers in or near the p53 locus. Two microsatellites, the AAAAT repeat located in intron 1 and the CA repeat located downstream of exon 11, together with the polymorphic sites at codons 72 and 213, were used for LOH determination. The microsatellites were analyzed according to the method of Ponteń et al. [23]. In short, microdissected tumor and corresponding normal tissue were amplified separately by a duplex single PCR, with one primer in each pair labeled with biotin, for 35 cycles. The PCR templates were purified by use of streptavidin-coated paramagnetic beads (Dynabeads® M-280 Streptavidin; Dynal AS) and denatured with NaOH. The eluted strands were neutralized with HCl and mixed with formamide. A fraction of each mixture was analyzed with Fragment ManagerTM (Pharmacia Biotech AB). The presence or absence of LOH was determined by comparing the signals of the two alleles in a tumor sample with the corresponding normal samples. In addition, LOH was determined in mutated/polymorphic samples by DNA sequence peak ratios. For mutated positions, a mutated peak signal markedly .50% directly indicated LOH. Thermocycling. PCR amplification of chromosomal DNA was performed as described previously [17]. Briefly, exons 4 to 9 of the human p53 gene and the HLA-DQB1 locus were amplified in a multiplex/nested configuration. The outer multiplex amplification was performed in one tube with 14 primers for 30 cycles by using both AmpliTaq® and Stoffel AmpliTaq® polymerase (Perkin-Elmer). After dilution (a 25-fold dilution for exons 4, 5, 7, and 8 and HLA and a 100-fold dilution for exons 6 and 9), the inner region-specific amplifications were performed (30 cycles). cDNA samples were amplified in four separate PCR reactions (38 cycles) with overlapping primer pairs covering exons 2 to 11, as described by Sjo¨gren et al. [15]. Thus, in both approaches, the p53 region was divided into multiple smaller fragments. All amplifications were performed by using a Perkin-Elmer 9600 thermocycler.

457

Clinical Chemistry 44, No. 3, 1998

performed in parallel) by estimation of the ratio between normal and altered DNA at polymorphic (mutated) positions. In the same manner, the ratio of mutated mRNA can be obtained from the RNA sequences.

Results Mutation analysis. One hundred breast cancer samples were compared by DNA sequencing of the p53 with genomic DNA microdissected from tumor lesions or RNA from homogenized tumor tissue as outlined in Fig. 1. All DNA and cDNA sequence alterations were reconfirmed by a repeated analysis of lysed cells or cDNA, respectively. In addition, LOH was performed on the microdissected material. All together, 26 samples with specific mutations were identified by the genomic DNA target assay; 23 of these alterations were located within the exons. Twenty-two mutations were detected by the RNA target assay (Table 1). In the sequences obtained from most tumors with mutations, the wild-type sequence could also be observed, which, because these tumors had loss of the wild-type allele, can be concluded to have originated from infiltrating or surrounding normal cells. Estimates of the mutated sequence to the background wild-type sequence (in a percentage of sequence peak signals) for all 26 samples are shown in Table 1.

Fig. 1. Schematic overview of the analysis methods for DNA- and mRNA-based sequencing of the p53 gene.

Direct solid-phase sequencing of p53 PCR products. One of the inner PCR primers for each fragment was labeled with biotin to facilitate solid-phase sequencing of PCR templates by the use of paramagnetic beads (Dynabeads M-280 Streptavidin; Dynal AS) for DNA analysis or plastic combs (AutoLoad® kit; Pharmacia Biotech AB) for RNA analysis, as solid support. Solid-phase sequencings were essentially performed according to the methods of Berg et al. [17] for DNA analysis and Sjo¨gren et al. [15] for RNA analysis and analyzed on an automated laser fluorescent sequencer apparatus (ALFTM; Pharmacia Biotech AB). The sequences were finally compared with the wild-type p53 sequence. All DNA and cDNA sequence alterations were reconfirmed by a repeated analysis of lysed cells or cDNA, respectively. For determination of mutation to wild-type signal in mutated samples, the PCR and following sequencing were repeated at least twice, the relative peak areas were estimated, and the mean value was reported. By this procedure, an indirect measurement of the residual amounts of normal cells in mutated samples can be obtained (if LOH analysis is

Missense mutations. Both assays identified 19 missense mutations. An important observation is that the mutated: wild-type ratios were consistently higher or equal in the RNA analysis than in the DNA analysis for all missense mutations (Table 1), as exemplified in Fig. 3. Thus, although we did not select for tumor cells by microdissection in the RNA assay, there was a higher tumor-specific signal for the RNA sample than for the DNA sample. All 19 tumors with missense mutations and increased mRNA concentrations stained positive with the p53 monoclonal antibody (Table 1). Truncating mutations. Six samples detected by the DNA assay, presented in Table 1, carried genomic mutations likely to yield truncated gene products (splice junctions and frame-shift or stop codon mutations); three of these were detected by using the mRNA assay, whereas two splice-site mutations and one stop codon mutation were not identified. Both splice-site mutations are located at the evolutionary conserved GT-junctions, which are known to be critical in the splicing process [24]. An additional alteration (sample 111), located 12 bp into the intron, as identified by the DNA assay, did not show any alteration in the mRNA analysis. The detected truncated mutations showed relatively less mutated mRNA in the sequence analysis compared with the mutated:wild-type ratio in the corresponding DNA sequence analysis. Although the exact transcription amounts in tumor and surrounding normal tissues remain to be determined, the steady-state concentrations of mRNA in the cells with p53 gene

458


Fig. 2. Stained tumor sections from patient 15 (A and B) and patient 18 (C and D) before (left) and after (right) removal of infiltrating normal cells by microdissection.

encoding truncated proteins seem not to be increased in the same way as the cells containing missense mutations. None of the tumors with truncated p53 product and relatively low mRNA concentrations stained with p53 monoclonal antibody (Table 1). Correlation with IHC and LOH. The tumors were analyzed by IHC using the p53 monoclonal antibody Pab 1801, which recognizes both native and mutated p53 proteins. Twenty-one tumors stained positive for the p53 protein, 19 of these had missense mutations, and 2 had wild-type sequence. None of the tumors with stop/splice-site or frameshift mutations stained positive. The overall results are presented in Table 1. The LOH analysis, performed on the microdissected samples, was primarily based on microsatellite analysis of two polymorphic repeat sequences, a penta-nucleotide (AAAAT) repeat located in intron 1 and a di-nucleotide (CA) repeat downstream of exon 11. In addition, sequencing data for samples heterozygous for codon 72 were used. In a few cases, the amount of contaminating normal cells in the tumor sample was too high for accurate determination. The different variables used for LOH analysis correlated well. When we used the independent assays for LOH, 81 of the 100 samples were informative. Of informative tumors with mutations, 21 of 22 had loss of the other allele (Table 1). Overall, 37 of 81 informative tumors retained both wild-type p53 alleles intact, and 22 retained one wild-type allele intact and deleted the other. Thus, 43 of 81 (53%) breast cancer tumors had LOH.

Control analysis. To confirm the previous observation that the RNA-based approach facilitates detection of missense mutations as a result of increased steady-state concentrations of mutated mRNA in tumors, control analyses were performed to exclude the possibility that previous comparative analyses were biased because of the slightly different cell populations used. A set of cancer tissue samples were analyzed, where DNA and RNA were coextracted, to achieve a direct comparison. Because no initial microdissection was performed, the contribution of normal cells was expected to be considerable. Sixteen samples (11 with missense mutations and 5 with truncating mutations) were analyzed by sequencing both mRNA and genomic DNA extracted from identical cell populations. Missense mutations. In the majority of samples with missense mutations (7 of 11), the relative mRNA concentrations were clearly increased (Table 2); one example is shown in Fig. 3. In the remaining four samples, the contribution of normal cells was negligible (both the RNA and the DNA assay showed close to a 100% signal for the mutated allele), making comparison of little value (Table 2). Truncating mutations. In contrast, the ratios of mutated to wild-type sequence for the truncating mutations were very similar in the DNA- and the RNA-based assay (Table 2), i.e., for this group of samples, no enhancement in peak signals of the mutated alleles was observed by using the RNA approach.

459


Table 1. A summary of data for the 26 (of 100) tumors, with mutations in the p53 gene. LOH Tumor

IHCa 0–9b

S-phasea (2/1)c

Aneuploidya (2/1)d

Codon mutated

Nucleotide change

Amino acid Mutated Mutated codon 72 change DNA (%)e RNA (%)e DNAf AAAAT

3 6 12 15 17 18 19 24 28 37 46 48 49 54 66 86 96 98 99 4 8 20 22 84 111 113

2 6 4 9 9 2 9 6 9 9 4 2 2 1 9 9 2 6 2 0 0 0 0 0 0 0

1 2 1 2 2 2 1 2 2 2 2 2 2 2 2 1 2 2 1 1 2 1 2 2 2 2

2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 2 2

237 157 237g 193 276 181 280 249 246 159 220 193 175 285 248 273 214 205 152 Intron 9 108 Intron 6 317 213 Intron 9 g 224

ATG-ATT GTC-GGC ATG-ATA CAT-CTT GCC-GGC CGC-CAC AGA-GGA AGG-AGT ATG-ACG GCC-GTC TAT-TGT CAT-CGT CGC-CAC GAG-AAG CGG-TGG CGT-TGT CAT-CGT TAT-TGT CCG-CTG gt-at 11 bp del gt-gg 1 bp ins CGA-TGA ctt-ctc GAG-TAG

Met-Ile Val-Gly Met-Ile His-Leu Ala-Gly Arg-His Arg-Gly Arg-Ser Met-Thr Ala-Val Tyr-Cys His-Arg Arg-His Glu-Lys Arg-Trp Arg-Cys His-Arg Tyr-Cys Pro-Leu Splice site Frameshift Splice site Frameshift Arg-stop Intron Glu-stop

15 60 100 50 20 35 50 75 30 60 30 25 80 80 30 95 95 10 30 j 30 85 75 75 15 100 50

30 85 100 90 50 85 85 75 100 90 50 60 100 95 50 100 100 100 75 wt 50 wt 60 30 j wt wt

n.i. n.i. n.i. 1 n.i. n.i. n.i. n.i. 1 n.i. n.i. n.i. n.i. n.i. n.i. n.d. n.i. n.i. n.i. n.i. n.i. n.i. n.i. n.i. n.d. n.i.

n.i. 1 1 1 1 1 n.i. 1 n.i. 1h n.i. n.i. 1 1 n.i. n.d. n.d. n.d. n.i. n.i. n.i. n.i. 1 n.d. n.d. 1

CA

n.i. 1 1 n.i. n.i. n.i. n.i. n.i. n.i. n.i. n.i. n.i. n.i. 1 n.i. n.d. n.i. 1 1 2 n.i. 1 1 1 n.d. 1

Type of Summary mutation

n.i. 1 1 1 1 1 n.i. 1 1 1 n.i. n.i. 1 1 1i 1 1 1 1 2 1 1 1 1 1 1

m m m m m m m m m m m m m m m m m m m t t t t t t t

a

Refers to analysis published by Sjo¨gren et al. [15 ]. The p53 IHC staining is given a number from 0 to 9 according to increasing strength and amount of nucleus stained; 1–9 is considered positive. c The proliferation is indicated by “1” for high S-phase and “2” for low S-phase. d 1, aneuploid tumors; 2, diploid tumors. e Mutated DNA or RNA corresponds to the ratio (in percent) of mutated nucleotide signal vs wild-type signal in the tumor sample (see Fig. 3); e.g., 100% indicates complete lack of wild-type signal, 50% equal amounts of wild-type and mutated signal, and only wild-type signal is indicated by “wt.” f The polymorphic codon 72 was used for LOH determination. g This mutation was also found in normal cells, as a heterozygote. h Microsatellite instability. i The polymorphic codon 213 indicated LOH. j Mutations that were first missed by the operators, but discovered through comparison of results. n.i., not informative; n.d., not determined; m, missense mutation; t, truncating mutation (stop, frameshift, and splice site). b

Discussion A detailed, sequence-based analysis of the p53 gene from 100 breast cancer tumor samples was performed, comparing the use of chromosomal DNA with extracted total RNA as templates. This has allowed evaluation of RNA sequencing vs DNA sequencing, along with analysis of relationships between immunohistology, steady-state concentrations of mRNA, LOH, and types of mutation in breast cancers. Generally, methods based on genomic DNA templates require great care to avoid normal cells in the tumor sample. A fine-tuned microdissection procedure [20, 23] has been established as a straightforward method to prepare cancer cells relatively free from normal cells. Obviously, the p53 alleles from both normal and cancer cells will be amplified with similar efficiencies. Because the sequencing analysis in

the end yields a compound signal of the two sequences, corresponding to the ratio of normal and cancer cells, mutation detection is virtually impossible in a sample that is ,25–30% tumor cells [25]. An advantage with the DNAbased method is that no extractions or precipitations have to be performed before the PCR, thus substantially decreasing the inherent problems of contamination. The method used here for DNA analysis covers only exons 4 to 9; thus, for full coverage of the p53 gene, exons 2, 3, 10, and 11 also need to be incorporated, requiring additional primers and optimization. The RNA method is, in contrast, a more efficient approach to cover the whole gene because the exons are joined in vivo by splicing. However, great care has to be taken to obtain fresh biopsies for the analysis, and an additional enzymatic step using reverse transcriptase must be performed.

460


Fig. 3. Comparison of sequence data of mutations obtained from DNA and RNA targets, respectively. Tumors are from patients 18 and 28 (see Table 1) and control sample 12A9 (see Table 2). The mutated nucleotide positions are indicated by arrows. Patients 18 and 28: The DNA assay shows a mixed sequence for both tumors, where 30% or less of the total peak signal corresponds to the mutated gene, showing a high ratio of normal cells in these samples (both tumors had loss of the wild-type allele). In contrast, the RNA sequence (bottom) for both samples shows almost 100% signal for the mutated nucleotide; i.e., the normal cells in these samples do not contribute any wild-type mRNA sequence. Control 12A9: Comparison of sequence data for sample 12A9 when using the exact same cell population. The DNA target sequencing shows that 10% of DNA in this sample is mutated, whereas the RNA target sequencing shows that 95% of the mRNA is mutated. Thus, the mutated p53 mRNA concentration is increased.

The strong correlation between IHC analysis and missense mutations suggests overexpression of the mutated p53 proteins in these cells (Table 1). Our data indicate that this overexpression might be a result of higher steadystate concentrations of mutated p53 mRNA as determined by the increase in mutated sequence signal when using RNA, rather than DNA, as a template. Additional control analysis of identical cell populations (Table 2) showed examples where 10 –30% of the DNA content was mutated but as much as 60 –100% of the corresponding RNA content was mutated. The higher concentrations of RNA in these tumors do not appear to associate with higher S-phase (Table 1). Thus, in contrast to previous observations [26, 27], we suggest that increased mRNA concentrations may contribute to the accumulation of p53 protein. Furthermore, all informative cases with a missense mutation (15 of 19) had LOH. Our study, therefore,

supports the view that inactivation of both alleles is essential for complete loss of p53 function [28, 29]. All tumors with mutations in the p53 gene coding for truncated gene products (splice junctions, frameshifts, and stop codon mutations) were IHC negative (Table 1). One of the nonsense mutations and two splice-site mutations could not be detected by the RNA analysis. This finding suggests lower steady-state concentrations of RNA encoding these mutations, compared with missense mutations, which is also supported by the control samples (Table 2). This observation is in agreement with the “nonsense-mRNA-decay” theory [30]. Furthermore, one additional alteration 12 bp into intron 9 was not detected by the RNA approach; this alteration, however, may not have affected the splicing. The overall lower mutant signal, relative to wild-type sequence, in the RNA than in the DNA approach can possibly explain the negative IHC staining.

461


Table 2. A summary of 16 control samples analyzed by coextraction of genomic DNA and mRNA followed by PCR and direct sequencing. Tumor

Codon mutated

Nucleotide change

Amino acid change

Mutated DNA (%)

Mutated RNA (%)

Type of mutationa

11E8 11G3 12A9 19 12D3 28 12G7 37 13B1 86 96 12B8 14D3 15B4 15C9 27D2

5:127 8:273 7:258 8:280 8:266 7:246 7:249 5:159 8:266 8:273 6:214 6:213 4:125 6:213 8:273 5:splicec

TCC-TTC CGT-CAT GAA-GGA AGA-GGA GGA-GTA ATG-ACG AGG-AGT GCC-GTC GGA-GTA CGT-TGT CAT-CGT CGA-TGA ACG-ACTb CGA-TGA 1AA ins ag-gg

Ser-Phe Arg-His Glu-Gly Arg-Gly Gly-Val Met-Thr Arg-Ser Ala-Val Gly-Val Arg-Cys His-Arg Arg-Stop 200-bp delb Arg-Stop Frameshift 21-bp delc

0 5 10 20 90 10 55 80 0 95 95 75 50 50 60 50

20 65 95 75 90 60 80 100 45 100 95 70 ;50b 55 40 65

m m m m m m m m m m m t t t t t

a

m, missense mutation; t, truncating mutation (stop, frameshift, and splice site). A silent mutation occurred at the last base of exon 4 and resulted in a 200-bp deletion of the mRNA (including one sequence primer site). An approximate estimation of the ratio of mutated to wild-type RNA in this sample was determined by comparing the intensity of ethidium bromide-stained PCR products, separated by agarose gel electrophoresis. c The splice-site mutation resulted in a 21-bp deletion in the corresponding mRNA, without causing a frameshift. b

These results, including mutation-type analysis, mRNA concentrations, LOH, and IHC, suggest that breast cancer patients can be grouped into at least four distinct groups with very similar p53 characteristics. A first group of patients has missense mutations, increased mRNA concentrations, loss of the other allele, and positive IHC. A second group of patients has truncated gene products, normal or lower mRNA concentrations, loss of the other allele, and negative IHC. A third group has one allele deleted (LOH) but the other gene intact, and a fourth group has both p53 alleles intact. The three groups with p53 gene alterations (i.e., mutation, LOH, or both) all have higher fractions of S-phase and aneuploidy (data not shown) than the fourth group with both p53 alleles intact. Obviously, these subgroups, which may have distinct clinical characteristics in respect to prognosis or therapy responses, cannot be characterized and identified by a single methodology. In this comparative study, we have focused primarily on the sequencing methods for characterization of exact genetic alterations in the p53 gene. Through evaluation of using either mRNA or genomic DNA templates, we conclude that an RNA target assay can take advantage of increased mRNA concentrations in breast cancer cells with p53 missense mutations, simplifying detection of tumor-specific sequences. Thus, the laborative microdissection step is not needed, and the analysis can be used on rather “crude” biopsies. However, the assay may not be fully sufficient for the detection of nonsense or splice-site mutations. This might give rise to results that are interpreted as false negatives, which is probably unacceptable in future routine analyses, although the exact clinical implications for these mutations

still are unknown. The main advantage of the DNA-based assay is that, when combined with microdissection, it can reliably identify all types of mutations and allows for parallel LOH analysis, which, when combined, can identify the members of each subgroup of cancer patients.

We are grateful to Anders Lindgren and Hans Nordgren for performing the IHC analysis and to Sarah Byding and Ingrid Blikstad for helpful assistance with RNA extraction and sequencing. This work was supported by grants from the Swedish Cancer Foundation, Pharmacia Biotech, and the Go¨ran Gustavsson’s Foundation.

References 1. Davidoff AM, Kerns BJM, Iglehart JD, Marks JR. Maintenance of p53 alterations throughout breast cancer progression. Cancer Res 1991;51:2605–10. 2. Greenblatt MS, Bennett WP, Hollstein M, Harris CC. Mutations in the p53 tumor suppressor gene: clues to cancer etiology and molecular pathogenesis. Cancer Res 1994;54:4855–78. 3. Coles C, Thompson AM, Elder PA, Cohen BB, Mackenzie IM, Cranston G, et al. Evidence implicating at least two genes on chromosome 17p in breast carcinogenesis. Lancet 1990;336: 761–3. 4. Davidoff AM, Humphrey PA, Iglehart JD, Marks JR. Genetic basis for p53 overexpression in human breast cancer. Proc Natl Acad Sci U S A 1991;88:5006 –10. 5. Singh S, Simon M, Meybohm I, Jantke I, Jonat W, Maass H, Goedde HW. Human breast cancer: frequent p53 allele loss and protein overexpression. Hum Genet 1993;90:635– 40. 6. Allred DC, Clark GM, Elledge R, Fuqua SA, Brown RW, Chamness GC, et al. Association of p53 protein expression with tumor cell

462

7.

8.

9.

10. 11. 12.

13.

14.

15.

16.

17.


proliferation rate and clinical outcome in node-negative breast cancer. J Natl Cancer Inst 1993;85:200 – 6. Thor AD, Moore DH II, Edgerton SM, Kawasaki ES, Reihsaus E, Lynch HT, et al. Accumulation of p53 tumor suppressor gene protein: an independent marker of prognosis in breast cancers. J Natl Cancer Inst 1992;84:845–55. Isola J, Visakorp T, Holli K, Kallioniemi OP. Association of overexpression of tumor suppressor protein p53 with rapid cell proliferation and poor prognosis in node-negative breast cancer patients. J Natl Cancer Inst 1992;84:1109 –14. Iggo R, Gatter K, Bartek J, Lane D, Harris AL. Increased expression of mutant forms of p53 oncogene in primary lung cancer. Lancet 1990;335:675–9. Dowell, SP, Hall PA. The clinical relevance of the p53 tumour suppressor gene. Cytopathology 1994;5:133– 45. Lane DP. p53 human cancers. Br Med Bull 1994;50:582–99. Healy E, Reynolds NJ, Smith MD, Campbell C, Farr PM, Rees JL. Dissociation of erythema and p53 protein expression in human skin following UVB irradiation, and induction of p53 protein and mRNA following application of skin irritants. J Invest Dermatol 1994;103:493–9. MacGeoch C, Turner G, Bobrow LG, Barnes DM, Bishop DT, Spurr NK. Heterogeneity in Li–Fraumeni families: p53 mutation analysis and immunohistochemical staining. J Med Genet 1995;32:186 – 90. Jansson T, Ingana¨s M, Sjo¨gren S, Norberg T, Lindgren A, Holmberg L, Bergh J. p53 status predicts survival in breast cancer patients treated with or without postoperative radiotherapy: a novel hypothesis based on clinical findings. J Clin Oncol 1995;13:2745–51. Sjo¨gren S, Ingana¨s M, Norberg T, Lindgren A, Nordgren H, Holmberg L, Bergh J. The p53 gene in breast cancer: prognostic value of complementary DNA sequencing versus immunohistochemistry. J Natl Cancer Inst 1996;88:173– 82. Borresen AL, Andersen TI, Eyfjord JE, Cornelis RS, Thorlacius S, Borg A, et al. TP53 mutations and breast cancer prognosis: particularly poor survival rates for cases with mutations in the zinc-binding domains. Genes Chromosomes Cancer 1995;14: 71–5. Berg C, Hedrum A, Holmberg A, Ponteń F, Uhleń M, Lundeberg J. Direct solid-phase sequence analysis of the human p53 gene by use of multiplex polymerase chain reaction and a-thiotriphosphate nucleotides. Clin Chem 1995;41:1461– 6.

18. Gumerlock PH, Chi S-G, Shi X-B, Voeller HJ, Jacobson JW, Gelmann EP, et al. p53 abnormalities in primary prostate cancer: single-strand conformation analysis of complementary DNA in comparison with genomic DNA. J Natl Cancer Inst 1997;89:66 – 70. 19. Bergh J, Norberg T, Sjo¨gren S, Lindgren A, Holmberg L. Complete sequencing of the p53 gene provides prognostic information in breast cancer patients, particularly in relation to adjuvant therapy and radiotherapy. Nat Med 1995;1:1029 –34. 20. Hedrum A, Ponteń F, Ren ZP, Lundeberg J, Ponteń J, Uhleń M. Sequence-based analysis of the human p53 gene based on microdissection of tumor biopsy samples. Biotechniques 1994; 17:118 –29. 21. Chevillard S. A method for sequential extraction of RNA and DNA from the same sample, specially designed for a limited supply of biological material. Biotechniques 1993;15:22– 4. 22. Chomczynski P, Sacchi N. Single-step method of RNA isolation by acid guanidinium thiocyanate–phenol– chloroform extraction. Anal Biochem 1987;162:156 –9. 23. Ponteń F, Berg C, Ahmadian A, Niste´r M, Lundeberg J, Uhleń M, Ponteń J. Molecular pathology in basal cell cancer with p53 as a genetic marker. Oncogene 1997;15:1059 – 67. 24. Krawczak M, Reiss J, Cooper DN. The mutational spectrum of single base-pair substitutions in mRNA splice junctions of human genes: causes and consequences. Hum Genet 1992;90:41–54. 25. Leitner T, Halapi E, Scarlatti G, Rossi P, Albert J, Fenyo¨ EM, Uhleń M. Analysis of heterogenous viral populations by direct DNA sequencing. BioTechniques 1993;15:120 – 6. 26. Hall PA, Lane DP. p53 in tumour pathology: can we trust immunohistochemistry?–Revisited! [Editorial]. J Pathol 1994;172:1– 4. 27. Thompson AM, Anderson TJ, Condie A, Prosser J, Chetty U, Carter DC, et al. p53 allele losses, mutations and expression in breast cancer and their relationship to clinico-pathological parameters. Int J Cancer 1992;50:528 –32. 28. Frebourg T, Sadelain M, Ng YS, Kassel J, Friend SH. Equal transcription of wild-type and mutant p53 using bicistronic vectors results in the wild-type phenotype. Cancer Res 1994;54:878 – 81. 29. Levine AJ, Momand J, Finlay CA. The p53 tumour suppressor gene. Nature 1991;351:453– 6. 30. Zhang S, Ruiz-Echevarria MJ, Quan Y, Peltz SW. Identification and characterization of a sequence motif involved in nonsense-mediated mRNA decay. Mol Cell Biol 1995;15:2231– 44.