Serum Anti-p53 Antibodies and Prognosis of Patients With Small-Cell Lung Cancer

Background: Some patients with cancer develop antibodies against the p53 tumor suppressor protein. The presence of these antibodies in serum has been associated with the expression of mutant p53 by the tumor and in some studies with a poorer survival. Purpose: The goals of this study were to determine the prevalence of anti-p53 antibodies in the serum of patients with newly diagnosed small-cell lung cancer (SCLC) and to assess the clinical relevance of the presence of these antibodies in the serum, particularly their relationship with tumor response to treatment and with patient survival. Methods: In this prospective study, serum was obtained from 170 patients atthe time of diagnosis of SCLC who were to subsequently receive platinumor doxorubicin-based chemotherapy at any one of four hospitals in Barcelona, Spain, from October 1991 through June 1994. Normal human sera from blood bank donors (n = 50) served as controls. The presence of anti-p53 antibodies was determined by western blot analysis with the use of purified recombinant p53 protein. As of January 1996, 96.5% of the patients had been treated and observed in the study, for a median follow-up time of 33.5 months. Survival was estimated by the Kaplan-Meier method. Cox proportional hazards regression and unconditional logistic regression analyses were conducted. All P values resulted from two-sided tests. Results: Anti-p53 antibodies were detected in the serum of 27 (16%) of the 170 patients studied. None of 50 serum samples from normal individuals contained anti-p53 antibodies. Analysis of pretreatment clinical characteristics demonstrated that a weight loss of less than 5% ( P = .025), a serum lactic acid dehydrogenase (LDH) level of less than 450 U/L ( P = .002), and limited stage disease (i.e., tumor confined to one hemithorax, with local and regional lymph node positivity for tumor cells and/or ipsilateral pleural effusion allowed) ( P < .001) were associated with a statistically significant complete response to therapy. The presence of serum anti-p53 antibodies was not associated with clinical characteristics, such as age ( P = .622), functional status ( P = 1.0), disease stage ( P =.634), complete response to treatment ( P = .572), and survival ( P = .492) or with any laboratory parameters including known prognostic factors in SCLC, such as serum sodium or LDH concentration ( P values of .731 and .246, respectively). Conclusions and Implications: The presence of anti-p53 antibodies in the serum of patients with newly diagnosed SCLC was not associated with any clinical characteristics or prognostic markers, suggesting that, in this context, the measurement of antip53 antibodies is not a useful prognostic marker. [J Natl Cancer Inst 1997;89:381-5] Mutations of the p53 tumor suppressor gene (also known as TP53) are among the most common genetic alterations in human cancers ( 1 ). Mutant p53 protein is thus a relatively common tumor antigen, and antibodies against p53 have been reported to occur in patients with a variety of cancers ( 2 – 4 ). The presence of anti-p53 antibodies has been associated with the accumulation of mutant p53 proteins inthe tumor and are considered a marker for the presence of these mutations ( 5 , 6 ). However, anti-p53 antibodies recognize both wild-type and mutant p53 proteins and the main immunogenic epitopes are at the p53 protein termini, not at the core domain where the majority of p53 mutations occur ( 5 , 7 ). This finding raises the issue of whether these antibodies are clinically or physiologically relevant. Since the first report of anti-p53 antibodies in patients with breast cancer in 1982 ( 8 ), studies ( 5 , 9 ) have been aimed at determining the clinical and pathologic relevance of these antibodies. One difficulty in interpreting results has been the use of enzyme-linked immunosorbent assay (ELISA) and immunoprecipitation studies that may yield false-positive results, particularly when rigorous controls are not used ( 10 ). In this study, we examined the sera of 170 patients with small-cell lung cancer (SCLC) for the presence of anti-p53 antibodies and assessed the association between the presence of these antibodies and clinical characteristics, response to treatment, and survival. From October 1991 through June 1994, 196 patients with newly diagnosed, histologically con firmed SCLC who were to be treated with conventional chemotherapy at four hospitals in Barcelona, Spain, were accrued for inclusion in this study. The sera of all of the patients were obtained by their physicians at the time of tumor diagnosis and storedat -70°C. Twenty-six of these patients were not treated or received only symptomatic treatment and were not included in the study, for a total of 170 patients in the study. Patients were considered to have limited disease if computed tomography (CT) scan of the chest and CT scan or ultrasound examination of the abdomen indicated that the tumor was confined to one hemithorax, to mediastinal and/or bilateral supraclavicular lymph nodes, or there was an ipsilateral pleural effusion; otherwise, patients were considered to have extensive disease. The limited disease stage of these patients was supported by the absence of metastases on a bone scan (obtained in 153 patients), a CT scan of the head (54 patients), or a bone marrow biopsy (60 patients) ( 11 ). The diagnostic testing used for staging was determined by the patient's oncologist. All 170 patients received either standard platinum (55%)- or doxorubicin (45%)- based chemotherapy ( 12 ). Response to treatment was considered to be complete if all disease found at the initial staging was absent as judged by imaging repeated at the completion of all treatments. A follow-up rate of 96.5% was achieved as of January 1996, and 10% of the patients were alive at the end of the study; the smedian follow-up time for all patients alive at last follow-up was 33.5 months (14.3–52.2 months). for all patients, survival was calculated from the date of diagnosis until the date of death or last clinical examination. To avoid the pitfalls of the ELISA and immunoprecipitation studies, we performed western blot analysis with the use of purified recombinant p53 protein. The full-length wild-type p53 complementary DNA (cDNA) was isolated from pCMV-HSV ( 13 ), subcloned into the pET21b expression vector (Novagen, Inc., Madison, WI), and purified by elution of the fusion protein from a Histidine Binding Nickel resin column according to the manufacturer's instructions (Novagen, Inc.). Purified p53 protein (15 µg/lane) was resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to nitrocellulose ( 14 ). To optimize results for potentially low-titer sera, the amount of protein resolved per lane was determined by titration of normal human serum against increasing amounts of purified protein. The highest amount of purified protein that yielded acceptable background was 15 µg/lane. The nitrocellulose strips were blocked overnight in 5% dry milk (Nestle, Glendale, CA) in phosphate-buffered saline (PBS), then incubated with 20 m M Tris (pH 8.0), 200 m M NaCl, 1% bovine serum albumin, and 0.2% Triton X-100 (TBST buffer) plus 10% sheep serum for 2 hours at room temperature. After extensive washing with 0.1% Tween in PBS, the nitrocellulose strips were incubated with a patient's serum for 1 hour (1:125-1:1000 dilution), followed by incubation with a horseradish-peroxidase-conjugated sheep anti-human immunoglobulin G diluted 1:40 000 for 1 hour at room temperature. The reaction was detected by enhanced chemiluminescence (Amersham Life Sciences Inc., Arlington Heights, IL). Immunoblots of mock protein lysates, obtained from bacteria containing the pET21b expression vector without the p53 cDNA insert, served as negative controls. Sera were considered positive for p53 antibodies if the immunoblot reactivity at the appropriate molecular weight (53 kd) titrated through serial dilutions as noted above. All of the sera were examined in at least two experiments and results were evaluated by two investigators. Incubation with a mouse monoclonal antibody recognizing p53 (1:40 000 dilution) (clone DO-1; Santa Cruz Biotechnology, CA) served as a positive control. Normal human sera from blood bank donors (n = 50) served as negative controls. Frequencies in contingency tables were compared by use of Pearson's chi-squared test or Fisher's exact test when greater than or equal to 20% of the cells had expected counts of less than five. The Student's t test was used to assess the differences in age among patients with or without anti-p53 antibodies ( 15 ). Survival distributions were computed by Kaplan-Meier methods and were compared by the logrank test ( 16 ). Hazard ratios and 95% confidence intervals (CIs) were estimated by use of proportional hazards models. The assumption of proportional hazards was checked graphically for each variable ( 17 ). We calculated adjusted odds ratios (ORs) and 95% CIs associated with complete response by multiple unconditional logistic regression models ( 18 , 19 ). Results were considered significant at the two-sided P <.05 level. All analyses were conducted with the use of the statistical packages Statview 4.01 (Abacus Concepts, Berkeley, CA) and SPSS 4.0 (Statistical Product and Service Solutions, Chicago, IL) for Macintosh computers. Clinical characteristics of the patients included in the study are shown in the first column of Table 1 . One hundred sixty-one of the 170 patients were male (95%), reflecting the cultural smoking patterns of this age group in Spain. The mean age ± standard deviation of all patients was 61 years ± 10 years. Forty-one percent of the patients had limited disease and 57% had extensive disease. Disease stage was unknown in three (2%) of the patients. All of the patients received standard chemotherapy (55% platinum based and 45% doxorubicin based). Of the 70 patients with limited disease, 46 (66%) patients also received chest radiotherapy. Twenty-five percent of all patients achieved a complete response. Anti-p53 antibodies were found in the serum of 27 (16%) of the 170 patients. None of the sera from the 50 normal individuals were found to exhibit anti-p53 reactivity. Fig. 1 shows representative re results of western blot analysis. Although antibody titers varied, all seropositive patients had antibodies detectable at serum dilutions of 1:250. There were no statistically significant differences in the clinical characteristics, laboratory parameters, and survival among patients with or without anti-p53 antibodies, including known prognostic factors in SCLC, such as serum lactic acid dehydrogenase (LDH) and sodium levels ( Table 1 ; Fig. 2 ) ( 20-22 ). In a univariate analysis of pretreatment clinical characteristics, a weight loss of less than or equal to 5% ( P = .025), a serum LDH of less than or equal to 450U/L ( P = .002), and limited stage disease ( P <.001) were all found to be associated with complete response to therapy. A multivariate logistic regression analysis was applied to assess the association of p53 antibodies with complete response to treatment after adjusting for age and hospital and including those variables that were associated with complete response in the univariate analysis; no association was detected (OR = 1.22; 95% CI =0.42–3.54). Kaplan-Meier curves did not show survival differences between patients who were positive and those who were negative for the presence of anti-p53 antibodies ( P = .492) ( Fig. 2 ). A Cox proportional hazards regression analysis adjusting for the prognostic factors identified in the univariate analysis was performed to determine whether anti-p53 antibodies had any independent prognostic value for survival; there was no association between the presence of anti-p53 antibodies and survival ( Table 2 ). This study represents a large series of prospectively recruited patients with SCLC in whom the presence of anti-p53 antibodies was examined by the use of a western blot analysis of recombinant p53 protein. We found that 16% of the patients with SCLC develop antibodies against the p53 tumor suppressor protein. However, no association was found between the presence of these antibodies and clinical features of the patients. One limitation of this investigation is that, in the study population, 95% of the patients were male. It is possible that female patients could develop antibodies with different frequency or clinical associations. Prior studies ( 2 , 8 ) have shown that 5%–24% of the patients with breast cancer or lung cancer harbor antibodies against p53 proteins in their serum. Patients with other tumors, such as colon, ovarian, pancreatic, brain, and head and neck squamous cell carcinomas, have also been reported to develop anti-p53 antibodies ( 4 , 7 , 23 – 25 ). In patients with breast cancer, the presence of p53 antibodies has been negatively associated with the presence of hormone receptors and positively associated with poorer survival ( 2 , 9 ). By contrast, another study ( 26 ) demonstrated that anti-p53 antibodies were more frequently detected in patients with breast cancer whose tumors had a favorable prognosis. In a study of patients with SCLC, Winter et al. ( 5 ) detected anti-p53 antibodies in four of 40 patients but found no association between the presence of these antibodies and stage, sex, or survival. Subsequently, these authors ( 27 ) reported that 21 of 36 patients with SCLC had antibodies against autologous tumor cell antigens (including two patients with anti-p53 antibodies) and that the presence of these antibodies was associated with improved survival. In a more recent study ( 25 ) in which ELISA was used, the presence of anti-p53 antibodies was found to be statistically significantly associated with increased risk of relapse and death in patients with squamous cell carcinoma of the head and neck. Detection of serum antip53 antibodies by western blotting. Each lane contains approximately 15 µg of purified p53 protein that has been resolved by sodium dodecylsulfate-polyacrylamide gel electrophoresis and probed with patients' sera. N = normal serum; (+) = purified monoclonal mouse anti-human p53 antibody used as positive control; A, B, and C = sera from patients with small-cell lung cancer (SCLC) and anti-p53 antibodies, and D and E = sera from patients with SCLC without p53 antibodies. All sera tested at 1:125 dilution. In addition, serum A is shown at dilutions of 1:125(A1), 1:250 (A2), 1:500 (A3), and 1:1000 (A4). A study ( 10 ), however, compared the detection of anti-p53 antibodies by ELISA, immunoprecipitation, and western blotting. There was a high frequency of false-positive sera detected by ELISA and immunoprecipitation, which on western blotting were found to be due to antibodies against a 53-kd protein that was not p53. In the present series, this artifactual limitation was overcome by the use of highly purified recombinant wild-type p53 protein. The use of wild-type p53 protein is supported by studies ( 5 , 7 ) that demonstrate that mutant p53 protein does not contain antigenic epitopes that are not present on the wild-type protein. The 16% prevalence identified in our study is similar to that reported in other series that used immunoblotting ( 5 ). No association was observed between the presence of anti-p53 antibodies and response to treatment and survival, either in univariate or multivariate analysis. Additionally, a multivariate logistic regression analysis of the data in our study showed that there was no association between p53 antibody status and complete response to therapy after adjusting for other known prognostic factors in SCLC, such as weight loss, performance status, and serum LDH levels ( 20 – 22 , 28 ). In this study, we found that a pretreatment weight loss of less than or equal to 5%, a serum LDH value of less than or equal to 450 U/L, and limited stage disease were associated with complete response to therapy. All patients in our series were studied at the time of SCLC diagnosis, demonstrating that p53 antibodies can be detected relatively early in the course of the disease. Since none of 50 serum samples from normal individuals were positive for p53 antibodies, the finding of these antibodies in patients at high risk, such as heavy smokers, could indicate the presence of an underlying cancer, as others have suggested ( 29 , 30 ). Kaplan-Meier survival curves for patients with small-cell lung cancer who had or did not have detectable serum anti-p53 antibodies. Patients with cancer develop antibodies against tumor antigens ( 31 , 32 ). To be clinically relevant, the presence of these antibodies should be associated with clinical parameters, such as response to treatment or survival. In patients with SCLC, the clinical and immunologic significance of serum anti-p53 antibodies remains unclear.