Immunohistochemical vs Molecular Biology Methods Complementary Techniques for Effective Screening ofp53 Alterations in Head and Neck Cancer

Transcription

1 ANATOMIC PATHOLOGY Original Article Complementary Techniques for Effective Screening ofp3 Alterations in Head and Neck Cancer ANNA CALZOLARI, PhD, ILARIA CHIARELLI, MSc, SIMONETTA BIANCHI, MD, LUCA MESSERINI, MD, ORESTE GALLO, MD, 3 BERARDINO PORFIRIO, PhD, AND PIER LUIGIMATTIUZ, MD The purpose of this study was to correlate p3 gene alterations and their expression in head and neck squamous cell carcinomas. Genomic p3 was amplified with the polymerase chain reaction (PCR) from formalin-fixed, paraffin-embedded tissues. Exons through were screened for mutations by means of nonisotopic single-strand conformation polymorphism (SSCP) analysis. p3 expression was detected by means of immunohistochemistry (IHC) with the p3 monoclonal antibody DO-. Twenty-three lesions (%) showed both SSCP variants and DO- immunostaining, whereas 3 (44%) demonstrated both SSCP normal patterns and negative staining. Twenty-five lesions (9%) were discordant, including IHC-positive (4%) and 3 SSCP-positive (%) lesions. However, discordant IHC-negative, The p3 tumor suppressor gene encodes a nuclear phosphoprotein. This functions as a transacting transcriptional regulator that controls the expression of a set of genes important in the regulation of the cell cycle and in triggering apoptosis after certain types of genomic damage. Mutation of the p3 tumor suppressor gene is the most common genetic alteration associated with cancers in humans. More than a thousand p3 mutations have been identified in human tumors. The majority of these mutations are found in the central amino acid portion of the protein. Biochemical studies have demonstrated that the core portion of p3 (residues to 9) folds into a compact structural domain From the ^Institute of Anatomic Pathology, Department of Clinical Physiopathology, Human Genetics Unit, and ^Institute of Otolaryngology Head and Neck Surgery, Florence University School of Medicine, Florence, Italy. Supported in part by grants from the Ministry of University (MURST), Rome, Italy; and by Fondazione Rulfo, Parma, Italy. Manuscript received May 9,99; revision accepted June,99. Address reprint requests to Dr Porfirio: Department of Clinical Physiopathology, Human Genetics Unit, University of Florence, Viale Morgagni,34, Florence, Italy. SSCP-positive lesions could have been easily interpreted after sequencing of the abnormal samples. Had these been screened with IHC only, all nonsense or frameshift p3 mutations would have been missed. IHC-positive, SSCP-negative lesions were interpreted in light of current models of p3 immunodetection. Had these been screened with SSCP or sequencing only, a sizeable number of tumors expressing p3 protein abnormalities would have been undetected. Therefore, simultaneous use of both methods increases the number of p3 abnormalities detected, and is warranted. (Key words: p3 tumor suppressor gene; Single-strand conformation polymorphism analysis; Immunohistochemistry; Head and neck squamous cell carcinoma) Am J Clin Pathol 99;:-. that contains the sequence-specific DNA binding activity of the protein. 3 The wild-type p3 protein has a short half-life and does not normally accumulate at levels high enough to be detected with immunohistochemistry (IHC). 4 A change in molecular configuration leads to stabilization of the protein; its half-life is greatly extended, thereby allowing detection at IHC. Although some studies have shown good correlation between IHC and the presence of mutations within the p3 gene detected at molecular analysis, - others have revealed varying degrees of discordance. 9 ' We correlated immunostaining by monoclonal antibody DO- with p3 gene mutations as assessed by means of single-strand conformational polymorphisms (SSCP) in a series of formalin-fixed, paraffinembedded head and neck squamous cell carcinomas. SSCP is a simple method for mutation screening that recognizes sequence variations through differences in the migration properties of single-stranded DNA in nondenaturing polyacrylamide gels. The efficiency of mutation detection with this technique has been reported to be as high as 9%. We show that DO- immunostaining and SSCP variants are correlated. Moreover, IHC-negative, SSCP-positive discor- Downloaded from by guest on November

2 ANATOMIC PATHOLOGY Article dant cases have been accounted for by sequencing, whereas IHC-positive, SSCP-negative discrepancies are interpreted in view of current models of p3 immunodetection. Patients MATERIALS AND METHODS Formalin-fixed, paraffin-embedded tumor specimens were obtained from consecutive, previously untreated head and neck cancers. All histologic slides stained with hematoxylin and eosin were reviewed by one pathologist (S.B.) to confirm the original diagnosis. The study included stage I or II and 3 stage III or IV head and neck squamous cell carcinomas of the larynx (n = ), oral cavity (n = ), oropharynx (n = ), and hypopharynx (n = ). Twelve tumors were poorly differentiated, 34 moderately differentiated, and 39 well differentiated. Sample Preparation Formalin-fixed, paraffin-embedded tissue sections to um thick were placed on standard microscope slides. Specimens were deparaffinized with xylene, rehydrated in serial graded (%, 9%, %, %) water-ethanol solution, rinsed in deionized water, and stained with hematoxylin and eosin. The samples were dehydrated again, but the slides were not fitted with a coverslip. The stained, unmounted sections were examined by microscopy. The portion of normal tissue from each specimen was identified and scraped away with a sterilized blade, leaving only tumor cells for DNA extraction. If discordant results between SSCP and IHC were obtained, different areas of the tumor were sampled to confirm the original data. DNA Isolation DNA from paraffin-embedded tumor sections was extracted by means of overnight incubation at C in an extraction buffer ( mmol/l potassium chloride, mmol/l tromethamine (tris) at ph.3,. mmol/l magnesium chloride,. mg/ml gelatin,.4%-np4,.4% polysorbate (Tween ), and. mg/ml proteinase K. The samples were boiled for minutes, and centrifuged for minutes at 4, rpm. One to ul of the supernatant was used in the polymerase chain reaction (PCR) mixture.. Polymerase Chain Reaction and Single-Strand Conformation Polymorphism Analysis Exons to of the p3 gene were subjected to PCR with intron primers. 3,4 Exon was amplified with two pairs of primers, which resulted in partially overlapping fragments. Amplification consisting of 3 cycles was carried out in ul total volume with. mmol/l MgCl buffer (Perkin-Elmer, Branchburg, NJ), umol/l of exon-flanking primer set, umol/l each of deoxyribonucleoside triphosphate, and. units of AmpliTaq (Perkin-Elmer). Temperature and time during the reaction cycles were 9 C for minute, C for 3 seconds, and C for 3 seconds. PCR products were heat denatured and subjected to SSCP analysis by electrophoresis on % polyacrylamide gel with % to % glycerol. Electrophoresis was carried out at 4 W for 4 hours at 4 C using the Pharmacia (Piscataway, NJ) apparatus. The gels were silver stained and dried on filter paper. All cases were amplified and run independently at least twice, with consistent results. To exclude the silent CGA/CGG dimorphism in codon 3, PCR products from samples showing SSCP abnormalities in exon were subjected to restriction analysis with Taq. Sequence Analysis PCR-amplified DNA strands were tested on 3% agarose gels (NuSieve; FMC Bioproducts, Rockland, Me), eluted from the gel by phenol extraction, and concentrated by ethanol precipitation. Purified products were sequenced on an automatic sequencer (Model 33A; Applied Biosystems, Foster City, Calif) using the dideoxy dye terminator method with the Taq polymerase and the PCR primers. Mutations were assessed on both strands. DO- Immunohistochemistry Five-micrometer paraffin-embedded sections were mounted on silane-coated slides and dried at C for 3 minutes. After dewaxing and blocking endogenous peroxidase, sections were rinsed in water and then placed in mmol/l of citrate acid at ph.. The sections were microwaved at W for minutes while covered with liquid. After microwave heating, the sections were transferred to phosphate-buffered saline solution. Specimens were stained with mouse anti-p3 protein antibody DO- (Dako, Carpenteria, Calif) at :4 in Downloaded from by guest on November AJCP y. iry 99

3 CALZOLARI ET AL 9 RESULTS IHC-positive moderate (++) and abundant (+++) staining was present in 3 cases (4%) (Fig ). Thirty- FIG. Poorly differentiated squamous cell carcinoma with intense staining for p3 protein in most tumor cell nuclei. six cases (4%) demonstrated SSCP band shifts suggestive of point mutations (Fig ). One sample was altered in both exon and. These results are consistent with the p3 mutation rate in most tumors in h u m a n s, including head and neck cancers. As shown in Table, the tumors in this series were stratified into distinct categories according to IHC and SSCP results. Twentythree lesions (%) were positive with both techniques. Twenty-seven tumors (3%) neither contained mutations detectable at SSCP nor stained for p3 protein, and an additional SSCP-negative specimens (%) demonstrated demonstrated fewer than % stained tumor cells. These lesions were tentatively scored as IHC negative, because multiple tumor sampling failed to produce a positive SSCP assay. Thus, total correlation between IHC and SSCP methods was attained in of cases (%) (P =.3, Fisher exact test). IHC and SSCP discordance was not related to tumor grade, size, or site (data not shown). Twelve lesions (4%) displayed moderate (++) or abundant (+++) p3 staining despite normal SSCP patterns. In contrast, lesions without nuclear reactivity and with focal (+) staining (%) showed SSCP shifts. These latter cases were subjected to sequence analysis, and a mutation was always found, confirming the good analytical p o w e r of SSCP (Table ). Ten mutations were either nonsense or frameshift alterations, thus generating a truncated protein. Two were base substitutions, leading to conservative amino acid changes (Arg to His, and Thr to Asn), whereas the remaining base substitution was of the n o n c o n s e r v a t i v e t y p e (Arg to Gly) a n d involved codon. TABLE. RELATIONSHIP BETWEEN DO- REACTIVITY AND SSCP ANALYSIS BY EXON IN HEAD AND NECK SQUAMOUS CELL CARCINOMAS DO- Immunoh 'stochemistry SSCP-negative (n=49) 4 SSCP-positive (n=3) Exon + Exon Exon Exon P*.3 FIG. Silver-stained polyacrylamide gel demonstrates aberrant SSCP bands, implying mutation in amplicon exon. Lanes and, tumor DNA from different samples; lane 3, normal DNA; lane 4, nondenatured normal DNA. Arrowheads indicate aberrant bands. SSCP = single-strand conformation polymorphism. Degree of immunostaining: + = <%; ++ = % to %; +++ = >%. *P value calculated by Fisher exact test, x -.4; df- 3. *One IHC-negative sample demonstrated SSCP shifts in both exon and exon. Vol. No. Downloaded from by guest on November phosphate-buffered saline solution. This antibody recognizes both m u t a n t a n d w i l d - t y p e p3 protein. Primary antibody bound to antigen was detected with a standard streptavidin-biotin technique (LSAB kit; Dako) and visualized with diaminobenzidine. A light hematoxylin nuclear counterstain was used. Control slides without the primary antibody were normal in all cases. Immunostaining was graded as follows: +, less than %; ++, % to %, and +++ = more than % stained cells.

4 ANATOMIC PATHOLOGY Article DISCUSSION Malignant tumors result from an accumulation of genetic alterations that lead to overexpression of oncogenes and inactivation of tumor suppressor genes. The consequences of these changes in cells progressing to malignancy are the loss of negative regulation of cellular proliferation or programmed cell death. Evidence thus far accumulated suggests that normal p3 protein is a negative regulator of cell growth, possibly playing a role in maintaining genomic integrity by arresting cells with DNA damage at the Gl/S boundary of the cell cycle. Mutations of the p3 gene are the most common genetic alterations associated with cancers in humans. An important consequence of these various mutations is posttranslational stabilization of the altered protein, leading to elevated p3 levels in tumor cells. Because of its short half-life, wildtype p3 protein is almost undetectable at IHC, and its positive immunoreactivity is considered synonymous with altered protein structure or function due to mutation or viral oncogene involvement., The presence of point mutations in the p3 gene is not the only cause of protein accumulation. Change in p3 half-life may also be caused by exposing normal cells to DNA-damaging agents, such as ultraviolet light. This critical observation has generated renewed interest and led to important new models of p3 protein function. 9 ' A cell in distress usually reacts by activation of biochemical pathways controlled by the p3 gene. The increase in p3 protein level is associated with an increase in transcription of p3-responsive genes and induction of growth arrest and apoptosis. The precise mechanism by which DNA damage and other stimuli result in p3 stabilization is not yet elucidated, but may involve the action of other gene products. Discrepancies between IHC and SSCP results may be interpreted on the basis of the above discussion. IHCpositive, SSCP-negative tumors can be explained by the presence of either mutations outside the exons screened or mutant conformers migrating like wild-type strands. However, a nonmutational stabilization of p3 may well be operative. Distressed cells express high levels of wild-type protein, perhaps through interruption of the p3 degradative pathway. Moreover, some p3- dependent genes involved in cell cycle regulation may be altered and may indirectly cause p3 overexpression in cell nuclei. On the other hand, IHC-negative, SSCPpositive rumors may result from nonsense point mutations or from deletions, because such alterations lead either to premature cessation of protein synthesis or to generation of quite different protein products, making IHC detection impossible. Our sequencing results confirmed this expectation; all but three mutations identified in discordant IHC-negative, SSCP-positive cases were single-base deletions (frameshift mutations) or premature stop codons (see Table ). Further insights into the nature of the discrepancies between IHC and molecular biology methods have recently been gained from crystallographic dissection of the p3 protein structure. 3 IHC-negative, SSCP-positive tumors may be related to the presence of point mutations that do not cause p3 protein stabilization. The socalled p3 class I mutants affect residues that directly bind DNA. 3 DNA binding domains are formed by a loop sheet helix motif coded by sequences in exons and and by a loop motif coded by sequences in exon? A variety of commercially available anti-p3 monoclonal antibodies have been developed for frozen or paraffin-embedded sections. In particular, monoclonal antibody 4 detects an epitope between and 33 residue. 4 However this monoclonal antibody works only on frozen sections, making it unsuitable for retrospective studies. p3 class I mutants fail to bind to monoclonal antibody 4, which is thought to recognize an epitope usually accessible in the mutant forms, 4 because they do not entail protein unfolding and hence stabilization and IHC detection. It is interesting that p3 gene alterations other than nonsense or frameshift mutations found in our IHC-negative tumors included two missense base substitutions in exon, leading to the conservative amino acid changes Arg to His (both positively charged) and Thr to Asn tooth hydrophilic) at codons and, respectively. The latter mutation Case No. 4 3* * + 9 TABLE. SEQUENCE ANALYSIS OF IHC-NEGATIVE, SSCP-POSITIVE SAMPLES Exon Codon Nucleotide Change CtoT CtoA Del lopb GtoA Del A DelG InsT Del CTA DelT AtoG DelG Amino Acid Change Gin to stop Thr to Asn Arg to His In frame deletion Arg to Gly IHC = immunohistochemistry; SSCP = single-strand conformation polymorphism. *IHC-negative and double mutated. ^Demonstrated focal immunostaining (<% positive cells). Downloaded from by guest on November AJCP* 99

5 CALZOLARI ET AL was coupled with a three base-pair deletion in exon (codon 4) that did not cause disruption of the protein coding frame. Whether in cis- or trans- configuration, these changes were undetectable with DO- IHC. p3 class II mutants affect residues that do not directly make contact with DNA but may be critical in stabilizing the protein structure. On the other hand, both the loop helix motif coded by sequences in exons and and the loop motif contain residues that, if altered, cause substantial protein unfolding, making p3 accessible to monoclonal antibody 4. In our series, all but two tumors with SSCP alterations in exon demonstrated positive immunostaining. The two exceptions were frameshift mutations that resulted in truncated products. Among factors that contributed to discordant IHCpositive, SSCP-negative results in a few cases, we should include incomplete SSCP screening, because some mutations may also be found in exons 4 and 9. However, protein stabilization in these cases may have been mediated by viral oncoproteins or by endogenous products, leading to immunostaining of wild-type p3. Furthermore, the SSCP-negative tumors with focal (<% stained cells) reactivity we classified as IHC negative may in fact include a minor fraction of p3 mutated cells missed at SSCP analysis. Finally, it should be emphasized that mutations that affect certain functional domains may lead to different IHC patterns and help explain IHC-negative, SSCP-positive cases. We speculate that, whereas missense mutations in exon, most often class II, are readily detectable with IHC, those in exons and may fail to determine protein conformational changes because of spotting in the loop sheet helix domain that directly contacts DNA. In conclusion, we found strong correlation between the presence of p3 protein accumulation and SSCP alterations in a series of head and neck squamous cell carcinomas. In general, data gathered by either of the methods predicted the results obtained by the other technique. It appears clear that discordant IHC and SSCP tumors also should be considered to bear some p3 abnormalities. Therefore, by enhancing the detection power of p3 aberrations in human neoplasia, the combined use of both techniques in a clinical setting may assist in management of cancer in many respects. Acknowledgments: Drs Calzolari and Chiarelli contributed equally to this study, and both should be regarded as first author of this article. We thank F. Caroli for skillful help in sequencing. REFERENCES. Levine AJ, Momand J, Finlay CA. The p3 tumor suppressor gene. Nature. 99;3: Lane DP. p3: guardian of the genome. Nature. 99;3:-. 3. Cho Y, Gorina S, Jeffrey PD, Pavletich NP. Crystal structure of a p3 tumor suppressor DNA-complex: understanding tumorigenic mutations. Science. 994;: Porter PL, Gown AM, Kramp SG, Coltera MD. Widespread p3 overexpression in human malignant tumors: an immunohistochemical study using methacarn-fixed, paraffin-embedded tissue. Am ) Pathol. 99;4:4-3.. Jenkins JR, Rudge K, Chumakov P, Currie GA. The cellular oncogene p3 can be activated by mutagenesis. Nature. 9; 3:-.. Rodrigues NR, Rowan A, Smith MEF, et al. p3 mutations in colorectal cancer. Proc Natl Acad Sci USA. 99;:-9.. Davidoff AM, Humphrey PA, Iglehart JD, Marks JR. Genetic basis for p3 overexpression in human breast cancer. Proc Natl Acad Sci USA. 99;:-.. Marks JR, Davidoff AM, Kerns BJ, et al. Overexpression and mutation of p3 in epithelial ovarian cancer. Cancer Res. 99;: Melhem MF, Law JC, El-Ashmawy L, et al. Assessment of sensitivity and specificity of immunohistochemical staining of p3 in lung and head and neck cancers. Am } Pathol. 99;4:-.. Allred DC, Clark G, Elledge R, et al. Association of p3 protein expression with tumor cell proliferation rate and clinical outcome in node-negative breast cancer. / Natl Cancer Inst. 993;:-.. Orita M, Suzuky Y, Sekiya T, Hayashi K. Rapid and sensitive detection of point missense mutations and DNA polymorphisms using the polymerase chain reaction. Genomics. 99;:4-9.. Hayashi K, Yandell DW. How sensitive is PCR-SSCP? Hum Mutat. 993;: Sakai E, Tsuchida N. Most human squamous cell carcinomas in the oral cavity contain mutated p3 tumor-suppressor genes. Oncogene. 99;: Toguchida J, Yamaguchi T, Ritchie B, et al. Mutation spectrum of the p3 gene in bone and soft tissue sarcomas. Cancer Res. 99;: Donehover LA, Bradley A. The tumor suppressor p3. Biochim Biophys Acta. 993;:-.. Lane DP, Benchimol S. Oncogene or anti-oncogene? Genes Dev. 99;4:-.. Tan TH, Wallis J, Levine AJ. Identification of the p3 protein domain involved in the formation of the SV4 large T antigen p3 protein complex. / Virol. 9;9:4-3.. Maltzman W, Czyzyk L. UV irradiation stimulates levels of p3 cellular tumor antigen in non-transformed mouse cells. Mol Cell Biol. 99;4: Lu X, Lane DP. Differential induction of transcriptionally active p3 following UV or ionizing radiation: defects in chromosome instability syndromes. Cell. 993;:-.. Fritsche M, Haessler C, Brandner G. Induction of nuclear accumulation of the tumor-suppressor protein p3 by DNA damaging agents. Oncogene. 993;:3-3.. Lane DP. The regulation of p3 function: Steiner Award Lecture. Int I Cancer. 994;:3-.. Hall PA, McKee PH, Manage HD, Dover R, Lane DP. High levels of p3 protein in UV-irradiated normal human skin. Oncogene. 993;: Vogelstein B, Kinzler KW. X-rays strike p3 again. Nature. 994;3:4. 4. Gannon JV, Greaves R, Iggo R, Lane DP. Activating mutations in p3 produce a common conformational effect: a monoclonal antibody specific for the mutant form. EMBO /. 99;9: 9-.. Hurlimann J, Chaubert P, Benhattar J. p3 gene alterations and p3 protein accumulation in infiltrating ductal breast carcinomas: correlations between immunohistochemical and molecular biology techniques. Mod Pathol. 994;:43-4. Downloaded from by guest on November Vol. No.