as, in humans, there exist a number of fatal disorders in which peroxisomes fail to import PTS1-containing enzymes [12, 13]. Peroxisome biogenesis disorders ...
unit/l), and shrimp alkaline phosphatase (SAP) was from Promega ... and 11 l of these reagent solutions were used to perform a differ- ..... area variations in the corresponding iTRAQ channels because the ..... ADP,ATP carrier protein. Q26365.
In this article, we survey several LC-MS/MS strategies which improve ... antibody (mAb) Rituximab following pellet digestion. .... improve quantitation of intact.
dimensional gels. This type of study should be valuable to develop diagnostic tools for watching water resources and to evaluate sensitivity to cyanotoxins (3). REFERENCES. 1. Jacquet, C ., et al. (2004) Toxicon 43, 141-147. 2.Schulenberg, B., Aggele
Offprint requests to: Yat-Pang, Chau., Ph.D., Institute of Anatomy and. Cell Biology ..... products demonstrated increased NQO1 transcript levels in most tumors .... were used, one against the normal form of p53 protein. (Ab-5) and the ..... 12, 566-
method based on the use of Surveyor nuclease (Transgenomic,. Omaha ... p53 immunohistochemistry assay . ..... Mutation detection using Surveyor nucle-.
expresses luciferase under the control of the human p21/Waf promoter were kind ... constructed by cloning the respective cDNAs into the pGEX-4T1 prokaryotic ...
Jun 8, 2004 - Biochem J 281 (Pt 1):21â40. Wang LH, Strittmatter ... Yasui F, Matsugo S, Ishibashi M, Kajita T, Ezashi Y, Oomura Y, Kojo. S, Sasaki K (2002) ...
... of ligand free, an appropriate solution of the binding equation to obtain the dissociation constant requires that we determine and fit the fractional saturation as a function of the concentration of total L added. The solution of the equation for
Sep 29, 2009 - plasma and the human serum protein solution Biseko. ... These fluids may be supplemented with electrolytes, proteins, plasma expander,.
Nov 28, 2006 - trated with mRNPs containing AU-rich element (ARE)-binding proteins, such as HuR, TIA1, TTP, and FXR1, during stress conditions. It therefore ...
... of a simple ligandâ receptor interaction (http://www.panvera.com/tech/fpguide/FP7.pdf). ... plot, a plot of ligand bound/ligand free vs ligand free (Fig. 1C). 1.3.
Jul 18, 2012 - +1-613-533-2825; Fax: +1-613-533-6830. Received: 17 May 2012; ... histologic (H) scores (% positive tumor area x staining intensity 0â3) were calculated using trained algorithms. H scores for all ...... Judy-Anne Chapman provided the
express the green fluorescent protein (GFP) reporter gene under the control of p53-responsive p21 promoter (i.e., the. p53R-GFP reporter system). Xenografts of ...
3 De Novo Pharmaceuticals, Compass House, Vision Park, Histon, Cambridge, CB4 9ZR, UK. The design, synthesis and evaluation of 24 isoindolinones as potential inhibitors of the MDM2-p53 interaction is described. The most potent inhibitor NU8231 (ELISA
Oct 15, 2009 - IN CANINE MALIGNANT MAMMARY TUMOURS ... protein overexpression and histological type, tumour size, or regional lymph node status.
http://cancerres.aacrjournals.org/content/51/23_Part_1/6304. Access the most recent version of this article at: E-mail alerts related to this article or journal.
Nov 7, 2017 - Aarhus, Denmark, 6 Department of Oncology, Aarhus University Hospital, Aarhus, Denmark, ... 4 [email protected] Abstract.
Apr 15, 2018 - staining was observed in the abnormal tissues of the biopsy samples from .... With the total 95 samples examined, positive p53 protein im- munoreactivity .... mutation in esophageal hyperplasia, a state known to revert back to.
bladder cancer in benzidine-exposed workers in Shanghai continuously for 20 yr .... Ï2 test, and log-transformed average scanning integrals of PCR amplified ...
May 9, 2011 - cross two inbred strains that are widely divergent for the trait, ... nonredundant database, it also necessitated the development of a strategy for grouping similar proteins due to a large number of shared peptides. Of the 4563 identifi
Apr 8, 2017 - models of the p53-Mdm2 feedback loop by quantitative protein monitoring. .... proteins including Arf, Siah, and h-catenin in p53 regulation.
Jan 29, 2014 - Salivary Glands of Sjogren's-Like Disease NOD Mice. Treated by ... PLoS ONE 9(1): e87158. doi:10.1371/journal.pone.0087158. Editor: Eva ...
isotope chemical labeling and tandem mass spectrometry. Phosphoproteins ... as viable alternatives for quantitative phosphorylation analysis.3r6. MS methods ...
Int. J. Cancer (Pred. Oncol.): 79, 494–501 (1998) r 1998 Wiley-Liss, Inc.
Publication of the International Union Against Cancer Publication de l’Union Internationale Contre le Cancer
QUANTITATIVE ANALYSIS OF p53 PROTEIN IN NON-SMALL CELL LUNG CANCER AND ITS PROGNOSTIC VALUE Michael A. LEVESQUE1,2, Mario D’COSTA2,3, Ernest H. SPRATT4,5, Mohammad M. YAMAN4,5 and Eleftherios P. DIAMANDIS1,2* 1Department of Pathology and Laboratory Medicine, Mount Sinai Hospital, Toronto, Canada 2Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto, Canada 3Department of Laboratory Medicine, St. Joseph’s Health Centre, Toronto, Canada 4Department of Surgery, St. Joseph’s Health Centre, Toronto, Canada 5Department of Surgery, University of Toronto, Toronto, Canada Accumulation of mutant p53 protein occurs frequently in human malignancies, including 40–60% of non-small cell lung carcinomas. The implications of such p53 over-expression, usually assessed by immunohistochemical techniques, for the prognosis of lung cancer patients remain undetermined. In this study, we used a time-resolved immunofluorometric assay to measure p53 protein concentrations in extracts prepared from 86 primary non-small cell lung tumours and examined the associations between p53 protein levels (corrected for total protein) and other clinico-pathologic variables, including post-surgical disease-free and overall survival. Contingency tables analysed by x2 tests revealed no significant relationships between p53 status, defined by a median cut-off point, and patient gender, age, disease stage, histologic grade and type, lymph node extension, smoking history and administration of adjuvant chemotherapy or radiation. However, multivariate Cox proportional hazard regression analysis demonstrated a dose-response relationship between p53 concentration, expressed as a 4-level, quartile-divided variable, and increased risk of relapse (p 5 0.010) and death (p 5 0.016). Patients whose tumours contained p53 concentrations exceeding the median value had over 3-fold higher risk of relapse (p 5 0.002) and death (p 5 0.007) than those whose tumours had lower p53 concentrations. We also provide evidence suggesting that the impact of p53 on survival is greater in patients with squamous cell carcinoma than in those with adenocarcinoma. Although the latter finding needs confirmation, our results suggest that application of an immunoassay of p53 protein on non-small cell lung tumour extracts may identify patients at increased risk of unfavourable outcome. Int. J. Cancer (Pred. Oncol.) 79:494–501, 1998. r 1998 Wiley-Liss, Inc.
Lung carcinoma is a leading cause of premature death in most countries and one of several malignancies for which the etiologic role of cigarette smoking has been clearly demonstrated. Unlike small cell lung cancer, which is less prevalent and typically disseminated at diagnosis but initially responsive to primary chemotherapy and radiotherapy, non-small cell lung cancer (NSCLC) accounts for approximately 85% of all cases and presents in one-third of these as localized disease treated primarily by surgery (Mountain, 1986). The prognosis of patients with advanced NSCLC is generally poor, and for this reason, adjuvant systemic therapy is usually instituted. Among patients with early-stage disease, however, 40–50% have been shown to relapse within 5 years after undergoing potentially curative resection (Mountain, 1986). Additional prognostic markers would be useful, particularly to identify those patients managed by surgery alone, who are at increased risk of early relapse or death and for whom adjuvant therapy may be beneficial. Mutation of the p53 gene has been shown to occur at very high frequencies in the majority of human cancers, including 40–60% of patients with NSCLC (Chiba et al., 1990; Kishimoto et al., 1992). These mutations, which are predominantly mis-sense and accompanied by nuclear accumulation of conformationally altered mutant p53 protein, have been detected in pre-neoplastic lesions of the lung (Sozzi et al., 1992; Bennett et al., 1993), in carcinomas in situ (Bennett et al., 1993) and in both primary tumours and metastases from individuals with extensive disease (Marchetti et al., 1993;
Fontanini et al., 1994), suggesting that p53 alteration may be an early event in NSCLC progression. Moreover, the high percentage of G-to-T transversion mutations of p53 in lung tumours has been linked to exposure to carcinogens present in cigarette smoke (Suzuki et al., 1992). The cellular consequence of p53 inactivation is thought to be the impairment of a pathway whereby cells harbouring damaged DNA are growth-arrested (Kastan et al., 1992) or deleted by programmed cell death (Lowe et al., 1993). A possible manifestation of this impairment in NSCLC may be resistance to chemotherapeutic agents, the efficacy of which is based on ability to induce p53-dependent apoptosis (Rusch et al., 1995). However, whether p53 mutations predict shortened postoperative survival in these patients more accurately than the existing prognostic factors based on tumour morphologic characteristics remains undetermined, despite having been a subject of investigation for many years (McLaren et al., 1992; Quinlan et al., 1992; Mitsudomi et al., 1993; Lee et al., 1995; Passlick et al., 1995; Nishio et al., 1996). Until quite recently, most studies of the prognostic implications of p53 in NSCLC have focused on the detection of p53 protein accumulation in tumour tissue rather than on direct analysis of the p53 gene for structural alterations. Although the simple immunohistochemical techniques commonly employed for this purpose have demonstrated good concordance with the results of DNA sequencing (Top et al., 1995), the different antibodies, specimenprocesssing details and scoring criteria used for immunostaining in these studies have likely contributed to the discordant conclusions. Enzyme-linked immunosorbent assays (ELISAs) of p53 protein have also been described (Vojteˇsˇek et al., 1992; Hassapoglidou et al., 1993; Levesque et al., 1995; Thomas et al., 1997) and may be advantageous in terms of sensitivity and specificity for the quantitative assessment of p53 protein in extracts of tumour tissues. With one exception (Pappot et al., 1996), these immunoassays have generally not been used to study the relationship between p53 protein accumulation status and the duration of survival in patients with NSCLC. Another such method, developed in our laboratory, has been used to demonstrate unfavourable prognosis associated with p53 protein over-expression in patients with breast carcinomas (Levesque et al., 1998) and has been shown to yield p53 protein concentrations in extracts of lung tumours, in agreement with the findings of immunostaining of the same tissues performed in parallel (Levesque et al., 1997). Since follow-up information, including that related to survival, was not initially available in the latter study, the prognostic value of ELISA-detected p53 protein in lung cancer could not be evaluated. The purpose of the present study was therefore to determine the relationships between p53 Grant sponsor: St. Joseph’s Health Care Foundation; Grant number: 1993–3. *Correspondence to: Department of Pathology and Laboratory Medicine, Mount Sinai Hospital, 600 University Avenue, Toronto, Ontario, Canada M5G 1X5. Fax: (416) 586–8628. E-mail: [email protected] Received 17 February 1998; Revised 25 April 1998
ELISA-DETECTED p53 PROTEIN AND LUNG CANCER PROGNOSIS
protein concentration and risk of relapse and death in a subset of patients who were prospectively monitored for a median of 16 months and whose lung cancers were restricted to NSCLC histologic types.
PATIENTS AND METHODS
Lung cancer patients In this study, we investigated a cohort of 86 patients operated at St. Joseph’s Health Centre, Toronto, Canada, from July 1993 to March 1995 for treatment of primary NSCLC. The ethics and research committee at this institution had given approval. Excluded from the consecutive series were patients with multiple synchronous lung tumours or other primary malignancies metastatic to the lung, as well as patients whose lung tumours were of a noninvasive or small cell histologic type, were judged to be nonresectable or were of insufficient quantity for p53 protein analysis, and patients for whom follow-up information was unavailable. Surgery followed diagnosis of lung cancer by an average of 9 days and consisted of lobectomy (n 5 46), pneumonectomy (n 5 21), wedge resection (n 5 12), lobectomy combined with wedge resection (n 5 5) or biopsy obtained at bronchoscopy (n 5 2). Ages of these patients, 58 of whom were male and 28 female, ranged from 42 to 87 years; the median age was 67 years. Among the 76 patients for whom history of tobacco use was known, the majority (n 5 68) recounted smoking on average the equivalent of one pack of cigarettes per day for 45 years (range 5 10–100 pack-years), while 8 patients remained non-smokers. Other clinico-pathologic features for which the majority of tumours had been characterised at the time of surgery included pathologic stage classification according to the tumour-node-metastasis (TNM) scheme (Beahrs et al., 1992) and the histologic grade and type based on World Health Organization (1982) criteria. Fifty-two patients (61%) were in stage I disease, 16 (19%) were in stage II, 13 (15%) were in stage IIIA, 1 (1%) was in stage IIIB and 3 (4%) had stage IV lung cancer. Of the 3 patients with stage IV disease, metastases to the bone and liver were identified in 2 cases and 1 case, respectively. Welldifferentiated (G1) tumours comprised 7% (n 5 6) of the specimens for which grade was known, while 62% (n 5 51) were moderately differentiated (G2) and 30% (n 5 25) were poorly differentiated (G3). Undifferentiated (G4) tumours were not present in our series. Tumour stage and grade could not be assessed for 1 and 4 patients, respectively. Most specimens were revealed to be either adenocarcinomas (n 5 43, 50%) or squamous cell carcinomas (n 5 37, 43%); the remainder consisted of large cell carcinomas (n 5 5, 6%) and an adenosquamous carcinoma (n 5 1, 1%). Regional lymph node metastases were identified in 32 of the patients. For all patients, follow-up was performed by review of medical records at St. Joseph’s Health Centre. Extended follow-up was also obtained by a questionnaire sent to other institutions and private practitioners throughout southern Ontario involved in the postoperative care. Informative responses were obtained for 56% of the questionnaires mailed. Forty patients (46%) underwent lung cancer relapse and 22 (26%) died of their disease during follow-up, which ranged from 1 to 44 months (median 5 16 months). Disease-free survival time, defined as the interval between the dates of tumour resection and the first evidence of recurrent or metastatic disease, ranged from 1 to 30 months (median 5 9 months). The corresponding period from surgery to death (the overall survival time) ranged from 1 to 40 months and had a median of 11 months. In the disease-free survival analysis, deaths without evidence of recurrence were considered censored observations, as were deaths due to causes other than lung cancer in the overall survival analysis (see below). Of the 61 patients remaining alive at last follow-up, 18 had relapsed and 43 were in remission. Prior to surgery, all patients had been untreated for lung cancer. Post-operatively, 13 patients received radiotherapy, 18 were administered systemic chemotherapy and 2 received both treatment modalities.
Tissue extraction and immunofluorometric assay of p53 protein Immediately following surgical excision, lung tumour tissues were snap-frozen and examined histologically, by which representative portions were selected for subsequent analysis. The percentages of tumour cells in these specimens were judged to be low (1–33%), medium (34–66%) or high (67–100%) in 7, 34 and 45 cases, respectively. After storage at –80°C for no more than 2 months, approximately 200 mg of these tissues were pulverized to a fine powder on dry ice and combined with 1 ml of a cell lysis buffer [50 mM Tris, pH 8.0, containing 150 mM NaCl, 5 mM EDTA, 1% (v/v) NP-40 surfactant, 10 mg/l phenylmethylsulfonylfluoride and 1 mg/l each of aprotinin and leupeptin], incubated for 30 min on ice and centrifuged for 30 min at 15,000 g at 4°C to collect the supernatants. The crude cell lysates were assayed directly for p53 protein using immunofluorometry and for total protein content using a kit based on the bicinchoninic acid method (Pierce, Rockford, IL). A ‘‘sandwich-type’’ immunoassay of p53 protein, described in detail previously (Levesque et al., 1995), was used to measure p53 protein concentrations in the lung tumour extracts. This method employed 2 p53-specific immunoreagents, monoclonal DO-1 antibody immobilised to microtitre plates prior to sample addition and polyclonal CM-1 anti-serum (Novocastra, Newcastle-upon-Tyne, UK), added subsequently and detected by an alkaline phosphataseconjugated goat anti-rabbit antibody (Jackson ImmunoResearch, West Grove, PA). Enzyme activity was detected by monitoring the formation of fluorescent complexes between the dephosphorylated reaction product diflunisal and EDTA-chelated Tb31, using timeresolved fluorometry (Christopoulos and Diamandis, 1992) performed on a dedicated instrument (Cyberfluor-615 Immunoanalyzer; MDS Nordion, Kanata, Ontario, Canada). Assay calibrator solutions ranged in concentration from 0.15 to 75 ng/ml and were prepared by dilution of a lysate of Sf9 insect cells infected with a p53-expressing baculovirus (Levesque et al., 1995). The p53 protein concentration of this lysate had been determined by assay in parallel against purified, bacterially expressed p53 protein (Oncogene Research, Cambridge, MA). Concentrations of p53 protein in the lung tumour extracts exceeding the detection limit of approximately 0.04 ng/ml were divided by the total protein concentrations to adjust for differences in tissue mass and extraction efficiency. Samples and calibrators were run as duplicates and in parallel. Since the epitope recognised by DO-1 antibody is within the amino terminal domain shared by wild-type and the majority of mis-sense p53 mutants (Vojteˇsˇek et al., 1992), the immunoassay is able to detect both mutant and wild-type p53 protein. Clinical specimens were assayed for p53 protein without knowledge of the corresponding clinico-pathologic or survival information. Statistical analysis All statistical procedures, performed by SAS for Windows version 6.12 software (SAS Institute, Cary, NC), were nonparametric and based on 2-sided tests of significance. Relationships between p53 protein concentrations and other continuous variables, such as age, number of pack-years smoked and the number of months until recurrence, death and last follow-up, were examined by calculation of Spearman correlation coefficients (rs). These associations, as well as those between p53 and several categorical variables, were also investigated by x2 tests applied to contingency tables. In these tables, the distributions of p53-negative and p53-positive tumours (using a cut-off point equal to the median p53 concentration, 0.14 ng/mg) were compared between patients differing with respect to age (less than 66 years vs. greater than or equal to 66 years), pathologic stage (I–II vs. III–IV), histologic grade (G1–G2 vs. G3) and type (adenocarcinoma vs. squamous cell carcinoma vs. other), lymph node involvement (absent vs. present), smoking history (no vs. yes), post-operative treatment (not treated vs. treated with chemotherapy and/or radiotherapy) and whether each of the clinical end points was reached during follow-up (non-relapsed vs. relapsed and alive vs. dead of lung cancer). The prognostic roles of p53, alone and in conjunction with other
LEVESQUE ET AL.
clinico-pathologic factors, in determining disease-free and overall survival were evaluated by hazards ratios [(HRs) relative risk (RR) of relapse or death] and 95% confidence intervals, which were calculated from fitted Cox proportional hazards regression models. In the multivariate analysis, the regression models were adjusted for age, stage, grade, node status and smoking history classified by the criteria given above. p53 protein concentration was tested in the models separately as a continuous variable; as a quartile-divided, 4-level ordinal variable; and as a dichotomous variable categorised by the median percentile cut-off point. Univariate and multivariate regression models were also used to evaluate the association between p53 protein concentrations exceeding the median value and post-surgical survival in the subgroups of patients diagnosed with adenocarcinoma, squamous cell carcinoma, early-stage (I–II) disease and late-stage (III–IV) lung cancer. p53-by-histotype and p53-by-stage interactions in the univariate Cox models were evaluated separately in order to determine if the effects of p53 on outcome were mediated by either stage or histotype. Each of the other dichotomous variables, including age, smoking history, grade, stage and node status, was similarly tested for its impact on prognosis by Cox regression analysis. The median p53 cut-off was also used in survival curves, examining the relapse or death rates of the p53-negative and p53-positive patients, which were plotted by the method of Kaplan and Meier and compared using log-rank statistics. RESULTS
Immunoassay of the 86 NSCLC tumour extracts revealed widely distributed p53 protein concentrations (mean 5 6.28 ng/ml, standard deviation 5 12.99 ng/ml, median 5 0.56 ng/ml, minimum 5 0.04 ng/ml, maximum 5 70.69 ng/ml), all of which exceeded the assay detection limit. Adjustment of these results for the total amount of protein extracted in each case led to a similarly positively skewed, somewhat bimodal distribution (Fig. 1), which had a mean of 1.22 ng/mg and a standard deviation of 2.24 ng/mg (range 0.005–10.97 ng/mg). The 25th percentile, median and 75th percentile of this distribution were 0.044 ng/mg, 0.14 ng/mg and 1.29 ng/mg, respectively. Protein-adjusted p53 concentrations were not significantly correlated with patient age (rs 5 0.14, p 5 0.198) or number of pack-years of cigarette smoking (rs 5 0.14, p 5 0.229) nor were they significantly associated with these variables and others, including gender, disease stage, histologic grade, histologic type, lymph node metastasis and whether or not adjuvant therapy was given, in contingency tables in which p53-positivity status was based on a median cut-off point (Table I). The same
FIGURE 1 – Frequency distribution of logarithmically transformed p53 protein concentrations in the 86 non-small cell lung tumour extracts. From left to right, dashed lines indicate the 25th, 50th and 75th percentiles of the distribution.
TABLE I – ASSOCIATIONS BETWEEN p53 STATUS1 AND OTHER CLINICOPATHOLOGIC VARIABLES Variable
Sex Female Male Age (years) ,66 $66 Stage3 I–II III–IV Histologic grade4 G1–G2 G3 Histologic type5 Adenocarcincoma Squamous cell carcinoma Other Lymph node metastasis No Yes Smoking history6 No Yes Post-operative treatment7 Not treated Treated Recurrence No Yes Death No Yes
p53 , median patients (%)
p53 $ median patients (%)
11 (39) 32 (55)
17 (61) 26 (45)
23 (60) 20 (42)
15 (40) 28 (58)
35 (51) 8 (47)
33 (49) 9 (53)
31 (54) 11 (44)
26 (46) 14 (56)
19 (44) 20 (54) 4 (67)
24 (56) 17 (46) 2 (33)
26 (48) 17 (53)
28 (52) 15 (47)
6 (75) 34 (50)
2 (25) 34 (50)
28 (53) 15 (45)
25 (47) 18 (55)
27 (59) 16 (40)
19 (41) 24 (60)
35 (55) 8 (36)
29 (45) 14 (64)
expression status based on median cut-off level of 0.14 ng/mg.–2p values calculated from x2 tests.–3Pathologic stage unknown for 1 patient.–4Histologic grade unknown for 4 patients.–5Other tumours were of large cell and adenosquamous histotypes.–6Smoking history unknown for 10 patients.–7Included treatment with chemotherapy or radiotherapy or both.
analysis performed using the 25th and 75th percentiles of the p53 protein distribution as cut-off points, as well as Mann-Whitney and Kruskal-Wallis tests, also did not yield statistically significant associations (data not shown), perhaps as a consequence of the small size of the study population. The relationships between p53 protein concentration and patient disease-free survival and overall survival were initially examined without considering the variable follow-up times, with which p53 is negatively correlated (rs 5 –0.22, p 5 0.039). Levels of p53 protein were thus also found to be negatively correlated with the length of the time interval from surgery to first recurrence (rs 5 –0.36, p 5 0.023) but not with those leading to patient death (rs 5 –0.11, p 5 0.586). A stronger association between p53 protein status, defined by a median cut-off point, and patient relapse was also suggested in contingency tables (Table I), though the tendency to have had p53-positive tumours of patients who relapsed or of those who subsequently died did not reach statistical significance. Selection of other quartiles as cut-off points in the contingency tables similarly did not lead to significant p values (data not shown). Univariate Cox regression analysis, however, in which p53 concentrations were used as a 4-level categorical variable or a dichotomous variable, revealed that increasing amounts of p53 protein in lung tumour extracts were associated with progressively reduced disease-free and overall survival probabilities (Table II). Although the use of p53 as a continuous variable in the regression analysis showed trends suggesting a dose-response relationship between p53 and risk of relapse (HR 5 1.01, p 5 0.097) and death (HR 5 1.01, p 5 0.050), these relationships were also evident from the analysis presented in the table in which patients were divided into 4 groups by the 3 quartiles of the frequency distribution of p53
ELISA-DETECTED p53 PROTEIN AND LUNG CANCER PROGNOSIS
TABLE II – ASSOCIATIONS BETWEEN p53 STATUS AND DISEASE-FREE AND OVERALL SURVIVAL p53 status
Disease-free survival HR1
Univariate analysis (n 5 86) Based on quartiles3 1st quartile 1.00 2nd quartile 0.93 3rd quartile 2.05 4th quartile 2.09 p value for trend Based on median value Negative 1.00 Positive 2.19 4 Multivariate analysis (n 5 75) 3 Based on quartile 1st quartile 1.00 2nd quartile 1.27 3rd quartile 3.74 4th quartile 4.00 p value for trend Based on median value Negative 1.00 Positive 3.02
Overall survival p value
1.00 0.83 1.55 3.31
0.35–2.48 0.83–5.06 0.85–5.19
0.20–3.37 0.41–5.76 1.00–10.97
0.021 1.00 2.47 1.00 0.97 2.16 17.16
0.47–3.48 1.25–11.17 1.18–13.53
0.21–4.45 0.51–9.18 2.12–139.14
0.016 1.00 3.72
is hazard ratio estimated by the Cox proportional hazards regression is the confidence interval of the estimated HR.–3Estimated HR for second, third and fourth quartiles relative to first quartile 4 are given.– Included in the multivariate models were patient age, stage, histologic grade, lymph node status and smoking history.
protein concentrations. The simpler division of patients into p53-negative and p53-positive groups of equal size by merging the 2 lower and 2 upper quartile-based groups demonstrated over 2-fold increased risks of relapse and death in p53-positive compared to p53-negative patients. These differences in survival rate between the median-dichotomised p53-negative and p53-positive patients were also shown by Kaplan-Meier survival plots (Fig. 2). Adjustment for the effects of age, stage, grade, node status and smoking history in multivariate Cox models of disease-free and overall survival revealed that p53 protein concentrations elevated above the median indicated significantly reduced survival independently of these other factors (Table II). Inclusion of treatment in the same multivariate models resulted in p53 losing its significance for overall survival (RR 5 2.33, p 5 0.12) but not for disease-free survival (RR 5 2.80, p 5 0.009). The magnitude of the increased risk of relapse or death conferred by p53-positivity in multivariate analysis was comparable to that associated with stage III–IV disease but of less statistical significance than the risks associated with lymph node metastases (Table III). In contrast, patient age, smoking history and histologic grade dichotomised as above were of no prognostic value in our series of 86 patients. Since the relationships demonstrated between p53-positivity and shortened disease-free and overall survival may have been restricted to patients whose tumours differed with respect to histologic type or anatomic extent, the risk of each outcome associated with p53-positivity was examined separately in subgroups defined by histotype and stage (Table IV). For patients with adenocarcinoma of the lung, the increased hazards of relapse and death estimated from the models for individuals with p53-positive tumour extracts were not significantly different from those for p53-negative patients. However, the more than 4-fold elevated risks of both end points for patients with p53-positive squamous cell carcinomas were of greater statistical significance. Although p53-positivity did confer an increased likelihood of relapse and death for both early-stage (I–II) and late-stage (III–IV) patients, the results were of greater statistical significance in early-stage lung cancer. In the more homogeneous subgroup of patients with stage I disease (n 5 52), other trends similarly suggested that p53positivity was associated with increased risk of relapse (HR 5 2.21, p 5 0.12) and death (HR 5 4.33, p 5 0.19). In patients with squamous cell carcinoma or early-stage disease (Table IV), but not
in those with adenocarcinoma or in stages III or IV (data not shown), p53-positivity remained an indicator of unfavourable prognosis in multivariate regression models. These models, which adjusted the relationships between p53 and survival outcomes for the influences of age, stage and lymph node status, revealed that patients with p53-positive squamous cell tumours had an 11-fold increased risk for relapse and a 35-fold increased risk for death compared with p53-negative patients with the same histologic type. The greater effect of p53 upon survival in squamous cell carcinoma patients compared with patients with adenocarcinoma could not likely be explained by differences in power since there were no marked differences in the numbers and proportions of patients relapsing or dying between the 2 subgroups: 21 and 14 of 43 adenocarcinoma patients relapsed and died, respectively, whereas 16 and 8 of 37 squamous cell carcinoma patients relapsed and died, respectively. However, the results of formal tests of interaction did not support a difference in the prognostic value of p53 between the 2 histologic classifications since the p53-by-histotype interaction term was not statistically significant for either relapse (p 5 0.23) or death (p 5 0.83). p53-by-stage interaction was similarly not significant in univariate Cox models.
Point mutations and allelic loss of the p53 gene are among the most common genetic lesions described in NSCLC, exceeded only by deletions on chromosome 3p (Brauch et al., 1987). Abundant evidence has suggested that this relatively high mutation rate results from the selective growth of tumour cell clones in which p53 has been functionally inactivated (Sozzi et al., 1995). Since p53 mutation appears to be an early, although not necessarily an initiating or even a required, event in lung tumorigenesis (Sozzi et al., 1992; Bennett et al., 1993; Fontanini et al., 1994), it may lead to a tumour phenotype which is unrelated to the local or systemic spread of the disease, represented by the stage classification. Pathologic stage is currently the major determinant for the receipt of adjuvant chemotherapy or radiotherapy; patients in early-stage disease are usually spared such treatment. The need for supplementary prognostic factors in NSCLC arises from the large proportion of patients with early-stage tumours, especially those without
LEVESQUE ET AL.
FIGURE 2 – Kaplan-Meier analysis of disease-free survival (a) and overall survival (b) of the 86 non-small cell lung carcinoma patients using the median value as the cut-off point for p53-positivity. p values were calculated by log-rank tests.
regional lymph node involvement, who nevertheless relapse and for whom aggressive adjuvant treatment may be warranted (Mountain, 1986). While a number of other morphologic and molecular features of NSCLC have been proposed to provide accurate prognostic information on which to base treatment decisions, none has emerged as unequivocally more valuable than stage alone. Some studies have shown mutation of the p53 gene (Horio et al., 1993; Mitsudomi et al., 1993) or over-expression of the p53 protein (Quinlan et al., 1992) to be indicative of shorter survival of NSCLC patients, but others have been unable to demonstrate such associations (McLaren et al., 1992; Lee at al., 1995; Passlick et al., 1995; Nishio et al., 1996). Differences in statistical power resulting from variable population sizes, as well as inherent differences in the populations themselves, may underlie many of these discordant findings. However, also contributing to the lack of consensus among these studies may be the differing sensitivities and specificities of the methods used to ascertain p53 mutational status. The immunohistochemical detection of p53 protein accumulation in the nuclei of tumour cells, including those from NSCLCs, has generally been accepted as evidence of p53 gene mutation. It has been estimated, however, that approximately 10–20% of p53 point mutations occur outside the most frequently mutated exons 5 to 8 and are not accompanied by p53 protein over-expression (Casey et al., 1996). Non-sense mutations, as well as nucleotide deletions and insertions, similarly are not associated with stabilisa-
tion of p53 protein, which may also occur by non-mutational mechanisms. Given the imperfect concordance between genetic alterations of p53 and accumulation of its protein product and the relative insensitivity of indirect techniques of screening for p53 mutations, such as single-strand conformation polymorphism analysis, it has been suggested that complete sequencing of all 11 exons is required to detect all p53 sequence changes (Casey et al., 1996). In vitro functional assays of p53 cloned from tumour cells to assess the ability of p53 mutants to regulate gene transcription in a sequence-specific manner, though biologically relevant, remain inappropriate for prognostic studies involving large numbers of patients. In contrast, immunohistochemical methods of detecting p53 protein have been predominantly used to demonstrate associations between p53 abnormalities and survival of cancer patients. Compared to conventional immunostaining techniques, however, densitometric image analysis of stained tissues (Charpin et al., 1996) and ELISA-type analyses of tissue extracts (Vojteˇsˇek et al., 1992; Hassapoglidou et al., 1993; Levesque et al., 1995; Thomas et al., 1997) may offer improved reproducibility. In addition, immunoassays, in their most common configuration, may possess greater specificity due to their use of 2 p53-directed antibodies. Despite these potential advantages, the use of ELISAs of p53 protein has been largely restricted to serum analysis (Luo et al., 1994), possibly reflecting the firm establishment of p53 immunostaining as a research tool in many histology laboratories, the inability of ELISAs to localize p53 accumulation in the context of histologic features and the requirement of fresh-frozen tissue specimens. General concordance between the findings of immunoassay and immunohistochemical analyses of several tissues has been reported (Vojteˇsˇek et al., 1992). In a study of 91 lung carcinomas of various histologic types, we found a strong correlation (rs 5 0.65, p , 0.001) between the p53 protein concentrations determined by our immunofluorometric assay in extracts of fresh-frozen tissues and the immunohistochemical scores reflecting the proportion of stained malignant cells, intensity of staining and tumour cellularity of matched formalin-fixed, paraffin-embedded tissues (Levesque et al., 1997). Since there were cases in that study for which the 2 methods were in disagreement, we considered it possible that p53 ELISA might yield prognostic information different from that of p53 immunostaining. The present study includes 82 NSCLC specimens assessed previously by the 2 methods and utilizes essentially the same median cut-off point for p53-positivity, equal to 0.13 ng/mg and 0.14 ng/mg in the previous and present studies, respectively. Selection of the median value as the cut-off was arbitrary but reasonable given the bimodal frequency distribution of p53 protein concentrations (Fig. 1) and consistent with the range of p53-positivity rates (40–80%) reported by others (McLaren et al., 1992; Bennett et al., 1993; Fontanini et al., 1994; Harpole et al., 1995; Lee et al., 1995; Passlick et al., 1995; Top et al., 1995; Nishio et al., 1996). Furthermore, we have shown that a small number of histologically normal lung tissues displayed p53 contents no greater than 0.07 ng/mg (Levesque et al., 1997). Using a commercially available pantropic p53 ELISA kit (Oncogene Research), Pappot et al. (1996) did not find a statistically significant difference in overall survival of 228 NSCLC patients whose tumour extracts had p53 levels below vs. above the median cut-off point. These workers observed, however, significantly higher p53 protein concentrations in squamous cell carcinomas than in adenocarcinomas, though the prognostic values of p53 within these subgroups were similiar, as well as higher p53 levels in patients who were older and who had large, late-stage tumours. Several other studies have also shown a slightly greater percentage of squamous cell carcinomas to exhibit p53 immunostaining than adenocarcinomas (Mitsudomi et al., 1993; Lee et al., 1995; Nishio et al., 1996), a relationship not consistent with our finding of a trend suggesting higher p53 levels in adenocarcinomas compared to squamous cell carcinomas. This latter relationship, however, was not statistically significant and should be confirmed in further studies of ELISA-detected p53 protein and NSCLC patient progno-
ELISA-DETECTED p53 PROTEIN AND LUNG CANCER PROGNOSIS
TABLE III – ASSOCIATIONS BETWEEN STATUS OF OTHER CLINICO-PATHOLOGIC VARIABLES AND PATIENT SURVIVAL Variable status
Disease-free survival HR1
Univariate analysis (n 5 86) Age (years) $66 vs. ,66 1.17 Smoking history3 Yes vs. no 0.81 Histologic grade4 G3 vs. G1–G2 1.15 Stage5 III–IV vs. I–II 2.63 Lymph node status Positive vs. negative 3.10 Multivariate analysis (n 5 75) Stage5 III–IV vs. I–II 3.18 Lymph node status6 Positive vs. negative 4.32
1HR is hazard ratio estimated by the Cox proportional hazards regression model.–2CI is the confidence interval of the estimated NR.–3Smoking history unknown for 10 patients.–4Histologic grade unknown for 4 patients.–5Pathological stage unknown for 1 patient.–6Multivariate models were adjusted for age, smoking history and p53 status.
TABLE IV – ASSOCIATIONS BETWEEN p53 STATUS1 AND SURVIVAL IN PATIENT SUBGROUPS p53 status
Disease-free survival HR2
Overall survival HR2
Univariate analysis Adenocarcinoma patients (n 5 43) Positive vs. negative 1.62 0.68–3.88 0.280 2.07 0.69–6.21 Squamous cell carcinoma patients (n 5 37) Positive vs. negative 4.41 1.40–13.88 0.011 4.64 0.91–23.73 Patients in stage I or II (n 5 68) Positive vs. negative 3.02 1.36–6.72 0.007 3.05 0.93–10.02 Patients in stage III or IV (n 5 17) Positive vs. negative 2.97 0.69–12.77 0.14 4.42 0.81–23.97 Multivariate analysis Squamous cell carcinoma patients (n 5 37)4 Positive vs. negative 11.68 2.89–47.20 ,0.001 35.69 2.85–447.86 Patients in stage I or II (n 5 68)5 Positive vs. negative 4.82 1.98–11.76 ,0.001 5.11 1.43–18.21 1p53
0.20 0.066 0.066 0.085 0.006 0.012
expression status based on median cut-off level of 0.14 is hazard ratio estimated by the Cox proportional hazards regression model.–3CI is the confidence interval of the estimated HR.– models were adjusted for patient age, stage and lymph node status.–5Mulitvariate models were adjusted for patient age and lymph node status.
sis. Although a few studies have reported p53 accumulation in NSCLC to be unrelated to any other clinico-pathologic variable (Chiba et al., 1990), a relationship between p53 abnormalities and lifetime exposure to cigarette smoke has been a finding shared by a large number of studies (Suzuki et al., 1992; Nishio et al., 1996). In survival analysis, levels of p53 protein exceeding the median of our series of 86 extracts of NSCLC tissues were also associated with reduced disease-free and overall survival. Moreover, we presented evidence that progressively increasing p53 concentrations may be associated with correspondingly increasing risks of relapse or death. That our findings differed from those reported by the only other study of the prognostic utility of ELISA-detected p53 protein in NSCLC (Pappot et al., 1996) was surprising since the median p53 concentration (0.10 ng/mg) used in the latter study was comparable to our median value and the median ages and proportions of patients in the various stages and histologic types were similiar, though the duration of follow-up for the larger number of patients in that study was greater. Other factors possibly accounting for the different conclusions of the 2 studies might also be differences in post-operative treatment or other undefined demographic or clinical differences between the patient populations which might have modified the impact of p53 on survival.
Since our series of NSCLC patients was not adequately characterised in terms of post-operative chemotherapy and radiotherapy administration, further studies of larger patient samples with detailed adjuvant treatment profiles are needed to determine the prognostic impact of ELISA-determined p53 protein in patients receiving different post-operative treatments. In multivariate analysis, the increased risk of relapse and death associated with p53-positivity was also shown to be of much greater magnitude in patients with squamous cell carcinomas than in those with adenocarcinomas. Previous studies had also suggested interaction between histologic type and p53 accumulation status but in such a way that it was in adenocarcinoma patients and not in those with squamous cell carcinoma that p53 was a statistically significant prognostic indicator (Nishio et al., 1996; Fukuyama et al., 1997). Taken together with our finding of a slight trend of greater p53-positivity in adenocarcinomas, the direction of the subgroup differences in the prognostic effect of p53 was quite surprising. Statistically significant interaction between p53 and histotype could not be demonstrated in Cox models of either relapse or death, and it remains possible that differences between patients with adenocarcinoma and squamous cell carcinoma which
LEVESQUE ET AL.
were not included in the multivariate models might have had a confounding influence on the relationships between p53 and survival. In particular, differences between subgroups in terms of chemotherapeutic agents received post-operatively could have led to our findings. Although power calculations were not performed, the similar numbers of patients and events within the 2 subgroups did not imply differences in power as an explanation. In contrast to the apparent differences with respect to p53 prognostic value between the 2 most common NSCLC histologic groups, patients with p53-positive tumours had similarly elevated risks for both outcome events irrespective of whether they had early (I–II)- or late (III–IV)-stage disease. Only those in the early-stage group were found in multivariate analysis to relapse or die at rates related to p53 status, but this finding is likely due more to the small number of patients in the late-stage group than to any true interaction between p53 status and stage. When the analysis was restricted to patients in stage I, who constituted 60% of the cohort, the same trends for poor survival associated with p53-positivity were evident. Other workers have also demonstrated p53 alteration to be indicative of poor prognosis in stage I or II NSCLC (Quinlan et al., 1992; Harpole et al., 1995) but not in stages III or IV (Passlick et al., 1995), while Mitsudomi et al. (1993) found p53 to be related to survival only in late-stage patients. The lack of consensus evidently points to a need for further investigation of larger, well-characterised patient series assessed for p53 abnormalities by reproducible techniques.
In this study, we have shown that the quantitative assay of p53 protein by a simple ELISA applied to extracts prepared from surgically resected NSCLCs may identify patients at increased risk of unfavourable outcome. We have also provided evidence that the utility of p53 quantification may be particularly relevant for the prognosis of patients with squamous cell carcinoma of the lung. Since our data confirm a number of previous reports of the prognostic value of p53 over-expression in this disease, we propose that the detection of p53 protein by ELISA may serve as an alternative to the more commonly employed immunohistochemistry. Given the small size and heterogeneity of our sample of patients, however, further studies are needed to unequivocally demonstrate the prognostic value of ELISA-detected p53 protein accumulation in patients with NSCLC.
This study was partially supported by a grant (1993–3) to M.D’C. from the St. Joseph’s Health Centre Foundation. The authors thank Dr. D.P. Lane for kindly providing the DO-1 hybridomas, Dr. T. Soussi for supplying the recombinant baculovirus expressing wild-type p53 protein, Dr. H. Yu for helpful discussions regarding the statistical analysis and Mr. D. Kang for technical assistance.
BEAHRS, O.H., HENSON, D.E., HUTTER, R.V.P. and KENNEDY, B.J. Manual for staging of cancer (3rd ed.), pp. 115–122, J.B. Lippincott, Philadelphia (1992). BENNETT, W.P. and 15 OTHERS, p53 protein accumulates frequently in early bronchial neoplasia. Cancer Res., 53, 4817–4822 (1993). BRAUCH, H., JOHNSON, B., HOVIS, J., YANO, T., GAZDAR, A., PETTENGILL, O., GRAZIANO, S., SORENSON, G., POIESZ, B., MINNA, J., LINEHAN, M. and ZBAR, B., Molecular analysis of the short arm of chromosome 3 in small-cell and non-small cell carcinoma of the lung. N. Engl. J. Med., 317, 1109–1113 (1987). CASEY, G., LOPEZ, M.E., RAMOS, J.C., PLUMMER, S.J., ARBOLEDA, M.J., SHAUGHNESSY, M., KARLAN, B. and SLAMON, D.J., DNA sequence analysis of exons 2 through 11 and immunohistochemical staining are required to detect all known p53 alterations in human malignancies. Oncogene, 13, 1971–1981 (1996). CHARPIN, C., GARCIA, S., BOUVIER, C., ANDRAC, L., DEVICTOR, B., LAVAUT, M.N., ALLASIA, C., BONNIER, P. and PIANA, L., Quantitative immunocytochemical assays on frozen sections of p53; correlation to the follow-up of patients with breast carcinomas. Amer. J. clin. Pathol., 106, 640–646 (1996). CHIBA, I. and 14 OTHERS, Mutations in the p53 gene are frequent in primary, resected non-small cell lung cancer. Oncogene, 5, 1603–1610 (1990). CHRISTOPOULOS, T.K. and DIAMANDIS, E.P., Enzymatically-amplified timeresolved fluorescence immunoassay with terbium chelates. Anal. Chem., 64, 243–346 (1992). FONTANINI, G., VIGNATI, S., BIGINI, D., MERLO, G.R., RIBECCHINI, A., ANGELETTI, C.A., BASOLO, F., PINGITORE, R. and BEVILACQUA, G., Human non-small cell lung cancer: p53 protein accumulation is an early event and persists during metastatic progression. J. Pathol., 174, 23–31 (1994). FUKUYAMA, Y, MITSUDOMI, T., ISHIDA, T., AKAZAWA, K. and SUGIMACHI, K., K-ras and p53 mutations are an independent unfavourable prognostic indicator in patients with non-small cell lung cancer. Brit. J. Cancer, 75, 1125–1130 (1997). HARPOLE, D.H., HERNDON, J.E., WOLFE, W.G., IGLEHART, J.D. and MARKS, J.R., A prognostic model of recurrence and death in stage I non-small cell lung cancer utilizing presentation, histopathology, and oncoprotein expression. Cancer Res., 55, 51–56 (1995). HASSAPOGLIDOU, S., DIAMANDIS, E.P. and SUTHERLAND, D.J.A., Quantification of p53 protein in tumour cell lines, breast tissues extracts and serum with time-resolved immunofluorometry. Oncogene, 8, 1501–1509 (1993). KASTAN, M.B., ZHAN, Q., EL-DEIRY, W.S., CARRIER, F., JACKS, T., WALSH, W.V., PLUNKETT, B.S., VOGELSTEIN, B. and FORNACE, A.J., A mammalian cell cycle checkpoint pathway utilizing p53 and GADD45 is defective in ataxia-telangiectasia. Cell, 71, 587–597 (1992).
KISHIMOTO, Y., MURAKAMI, Y., SHIRAISHI, M., HAYASHI, K. and SEKIYA, T., Aberrations of the p53 tumor suppressor gene in human non-small cell carcinomas of the lung. Cancer Res., 52, 4799–4804 (1992). LEE, J.-S., YOON, A., KALAPURAKAL, S.K., RO, J.Y., LEE, J.J., TU, N., HITTLEMAN, W.N. and HONG, W.K., Expression of p53 oncoprotein in non-small cell lung cancer: a favorable prognostic factor. J. clin. Oncol., 13, 1893–1903 (1995). LEVESQUE, M.A., D’COSTA, M., ANGELOPOULOU, K. and DIAMANDIS, E.P., Time-resolved immunofluorometric assay of p53 protein. Clin. Chem., 41, 1720–1729 (1995). LEVESQUE, M.A., KATSAROS, D., YU, H., GIAI, M., GENTA, F., ROAGNA, R., PONZONE, R., MASSOBRIO, M., SISMONDI, P. and DIAMANDIS, E.P., Immunofluorometrically-determined p53 accumulation as a prognostic indicator in Italian breast cancer patients. Int. J. Cancer (Pred. Oncol.) 79, 147–152 (1998). LEVESQUE, M.A., TADROSS, L., DIAMANDIS, E.P. and D’COSTA, M., Comparison of immunofluorometry and immunohistochemistry for the detection of p53 protein in lung cancer specimens. Amer. J. clin. Pathol., 107, 308–316 (1997). LOWE, S.W., SCHMITT, E.M., SMITH, S.W., OSBORNE, B.A. and JACKS, T., p53 is required for radiation-induced apoptosis in mouse thymocytes. Nature (Lond.), 362, 847–849 (1993). LUO, J.-C., ZEHAB, R., ANTTILA, S., RIDANPAA, M., HUSGAFVEL-PURSIAINEN, K., VAINIO, H., CARNEY, W., DEVIVO, I., MILLING, C. and BRANDT-RAUF, P.W., Detection of serum p53 protein in lung cancer patients. J. occup. Med., 36, 155–160 (1994). MARCHETTI, A., BUTTITA, F., MERLO, G., DIELLA, F., PELLEGRINI, S., PEPE, S., MACCHIARINI, P., CHELLA, A., ANGELETTI, C.A., CALLAHAN, R., BISTOCCHI, M. and SQUARTINI, F., p53 alterations in non-small cell lung cancers correlate with metastatic involvement of hilar and mediastinal lymph nodes. Cancer Res., 53, 2846–2851 (1993). MCLAREN, R., KUZU, I., DUNNILL, M., HARRIS, A., LANE, D. and GATTER, K.C., The relationship of p53 immunostaining to survival in carcinoma of the lung. Brit. J. Cancer, 66, 735–738 (1992). MITSUDOMI, T., OYAMA, T., KUSANO, T., OSAKI, T., NAKANISHI, R. and SHIRAKUSA, T., Mutations of the p53 gene as a predictor of poor prognosis in patients with non-small cell lung cancer. J. nat. Cancer Inst., 85, 2018– 2023 (1993). MOUNTAIN, C.F., A new international staging system for lung cancer. Chest, 89, 225–233 (1986). NISHIO, M., KOSHIKAWA, T., KUROISHI, T., SUYAMA, M., UCHIDA, K., TAKAGI, Y., WASHIMI, O., SUGIURA, T., ARIYOSHI, Y., TAKAHASHI, T., UEDA, R. and TAKAHASHI, T., Prognostic significance of abnormal p53 accumulation in primary, resected non-small cell lung cancers. J. clin. Oncol., 14, 497–502 (1996).
ELISA-DETECTED p53 PROTEIN AND LUNG CANCER PROGNOSIS
PAPPOT, H., FRANCIS, D., BRU¨ NNER, N., GRØNDAHL-HANSEN, J. and ØSTERLIND, K., p53 protein in non-small cell lung cancer as quantified by enzyme-linked immunosorbent assay: relation to prognosis. Clin. Cancer Res., 2, 155–160 (1996). PASSLICK, B., IZBICKI, J., HA¨ USSINGER, K., THETTER, O. and PANTEL, K., Immunohistochemical detection of p53 protein is not associated with a poor prognosis in non-small cell lung cancer. J. Thorac. Cardiovasc. Surg., 109, 1205–1211 (1995). QUINLAN, D.C., DAVIDSON, A.G., SUMMERS, C.L., WARDEN, H.E. and DOSHI, H.M., Accumulation of p53 protein correlates with a poor prognosis in human lung cancer. Cancer Res., 52, 4828–4831 (1992). RUSCH, V., KLIMSTRA, D., VENKATRAMAN, E., OLIVER, J., MARTINI, N., GRALLA, R., KRIS, M. and DMITROVSKY, E., Aberrant p53 expression predicts clinical resistance to cisplatin-based chemotherapy in locally advanced non-small cell lung cancer. Cancer Res., 55, 5038–5042 (1995). SOZZI, G., MIOZZO, M., DONGHI, R., PILOTTI, S., CARIANI, C.T., PASTORINO, U., DELLA PORTA, G. and PIEROTTI, M.A., Deletions of 17p and p53 mutations in preneoplastic lesions of the lung. Cancer Res., 52, 6079–6082 (1992). SOZZI, G., MIOZZO, M., PASTORINO, U., PILOTTI, S., DONGHI, R., GIAROLA, M., DE GREGORIO, L., MANENTI, G., RADICE, P., MINOLETTI, F. and DELLA
PORTA, G., Genetic evidence for an independent origin of multiple preneoplastic and neoplastic lung lesions. Cancer Res., 55, 135–140 (1995). SUZUKI, H., TAKAHASHI, T., KUROISHI, T., SUYAMA, M., ARIYOSHI, Y., TAKAHASHI, T. and UEDA, R., p53 mutations in non-small cell lung cancer in Japan: association between mutations and smoking. Cancer Res., 52, 734–736 (1992). THOMAS, M.D., MCINTOSH, G.G., ANDERSON, J.J., MCKENNA, D.M., JOHNSTONE, R., LENNARD, T.W.J., HORNE, C.H.W. and ANGUS, B., A novel quantitative immunoassay system for p53 using antibodies selected for optimum designation of p53 status. J. clin. Pathol., 50, 143–147 (1997). TOP, B., MOOI, W.J., KLAVER, S.G., BOERRIGTER, L., WISMAN, P., ELBERS, H.R.J., VISSER, S. and RODENHUIS, S., Comparative analysis of p53 gene mutations and protein accumulation in human non-small cell lung cancer. Int. J. Cancer (Pred. Oncol.), 64, 83–91 (1995). ´ RTEK, J., MIDGELY, C.A. and LANE, D.P., An immunochemiVOJTEˇ Sˇ EK, B., BA cal analysis of the human nuclear phosphoprotein p53. J. Immunol. Methods, 151, 237–244 (1992). WORLD HEALTH ORGANIZATION, Histological typing of lung tumors. Amer. J. clin. Pathol., 77, 123–136 (1982). WYNFORD-THOMAS, D., p53 in tumour pathology: can we trust immunocytochemistry? J. Pathol., 166, 329–330 (1992).