Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

p53 is the most mutated gene identified in human neoplasms (1). p53 acts as a tumor-suppressor gene and findings in several studies suggest that it also has an important role in control of the cell cycle (2). p53 has been shown to control an important cell-cycle checkpoint responsible for maintaining the integrity of the genome (3), and to be involved in apoptosis triggered under certain conditions (4, 5). It has been shown that wild-type p53 mediates arrest of the cell cycle in the G1 phase after sublethal DNA damage (6). It also appears that p53 is involved in transcriptional control (7, 8).

Several lung-cancer studies have shown a direct correlation between immunohistochemical detection of the p53 protein and p53 gene mutation. It is believed that immunocytochemical evidence for this protein is an indication of p53 gene mutation (9).

p21 was the first inhibitor of the cyclin-dependent kinase (cdk) complexes to be discovered, and is also known as CIP1 (14), SDI1 (15), mda-6 (16), and WAF1 (17). On examining the genomic structure of p21, it was found that its promoter contains two elements recognizable by p53. Binding of wild-type p53 to these regions was found to upregulate transcription of p21 (17), such that p21 is now considered to be a downstream target of p53. The pathway by which p53 directly inhibits DNA replication has recently been shown to involve at least two mechanisms. The p21 protein, transcriptionally unregulated by p53, can inhibit the kinase activities of the G1 cyclin/cdk complexes cdk4/cyclin D and cdk2/cyclin E, two complexes that are absolutely necessary for onset of the S phase of the cell cycle. If DNA damage occurs during replication, p21 can inhibit the cdk2/cyclin A complex, which is necessary for the cell to progress through S phase and is also closely associated with DNA replication (18). The p53-to-p21 pathway also inhibits DNA replication by p21 interaction with proliferating-cell nuclear antigen (PCNA). It has been shown that p21 can inhibit PCNA SV40 genome replication in vitro (19, 20). Recently, Li and coworkers (21) demonstrated that p21-PCNA interaction does not affect the DNA repair abilities of PCNA. Different lines of investigation recently gave evidence that p21 also can be induced by p53-independent pathways (22).

Lung cancer now has one of the worst prognostic potentials among human malignancies. Only about 30% of patients are candidates for radical lung resection, and only 30% of these patients are free of disease within 5 yr after surgery (25). New approaches to the management of this disease are urgently required to better define the profile of patients eligible to undergo curative surgical resection. In order to clarify critical events during bronchial carcinogenesis, and to evaluate a possible prognostic role for immunohistochemical detection of p21, we assessed the immunohistochemical expression of p21 protein in 60 surgically resected non-small-cell lung cancers (NSCLCs) that had previously been investigated for their p53 protein status (12).

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Immunohistochemistry

We assessed the immunohistochemical expression of p21 protein in 60 specimens consisting of 39 squamous-cell carcinomas and 21 adenocarcinomas obtained from a cohort of patients with surgically resected lung cancer in which the p53 status had previously been characterized. One specimen of that cohort was unavailable for examination.

Specimens were obtained only from patients who did not receive chemo- or radiotherapy prior to surgical resection. The histologic diagnoses and classifications of the tumors were based on the World Health Organization (WHO) criteria (26). The postsurgical pathologic tumor-nodal- metastatic (TNM) stage of each malignancy was determined according to the guidelines of the American Joint Committee on Cancer (27). One tumor was classified as Stage I, 52 tumors were classified as Stage II and seven tumors were classified as Stage IIIa. With regard to histologic grading, eight specimens were poorly differentiated, 42 specimens showed medium differentiation, and 10 specimens were highly differentiated.

Survival data were collected from hospital charts and from periodic interviews with patients and their relatives. The follow-up period for surviving patients was 48 mo. Patients dying of causes other than lung cancer during the follow-up period were excluded from the study.

Sections from each specimen were cut at a thickness of 3 to 5 µm, mounted on glass, and dried overnight at 37°C. All sections were deparaffinized in xylene, rehydrated through a graded alcohol series, and washed in phosphate-buffered saline (PBS). This buffer was used for all subsequent washes and for dilution of the antibodies.

Sections were heated twice in a microwave oven for 5 min each at 700 W in citrate buffer (pH 6), sequentially quenched in 0.5% hydrogen peroxide, and blocked with diluted 10% normal horse antimouse serum (Vector Laboratories, Burlingame, CA). A monoclonal antibody raised against p21 (Ab-1; Oncogene Science, Cambridge, MA) was used (dilution 1:100). The incubation time was 120 min at room temperature. After washing in PBS, the slides carrying the sections were incubated with diluted, biotinylated horse antimouse antibody (Vector Laboratories) for 30 min at room temperature. The slides were then processed by the avidin-biotin conjugate (ABC) method (Vector Laboratories) for 30 min at room temperature. Diaminobenzidine (DAB) was used as the final chromogen and hematoxylin as the nuclear counterstain. Negative controls for each tissue section were produced by leaving out the primary antibody. Because p21 protein is almost always immunohistochemically detectable, immunohistochemistry was also performed on 10 normal bronchial specimens in order to establish the normal p21 expression pattern of this tissue for use as a positive control.

All samples were processed under the same conditions. Three pathologists (A. B., A. G., and F. B.) evaluated the staining pattern of the protein separately, and scored it for the percentage of positive nuclei (0 = no positive staining; 1 = 1% to 30% positive cells; 2 = 30% to 60% positive cells; and 3 = more than 60% positive cells). The same criteria had been adopted for p53 scoring (12).

Statistical Analysis

We analyzed the study data statistically to verify any relationship between histologic type, histologic grading, tumor stage, or TNM status and protein-expression level. These analyses were assessed with Spearman's rank correlation test.

Patient-survival data were used to determine any correlation between p21 expression level and overall survival time. Survival curves were constructed through a Kaplan-Meier analysis. Statistical significance of these data was measured with analysis of variance (ANOVA) and chi-square testing (27).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Immunoreactivity for p21 was found in both normal and neoplastic tissues. p21 immunostaining was always nuclear, with a low to absent background. p21 expression was detected in all normal bronchial specimens, but it was restricted to the ciliated cells (the basal cells were unreactive), to a few isolated gland cells (Figure 1), and to Type II alveolocytes, ranging from 1% to 5% of the total cells.

View larger version (106K): [in this window] [in a new window]   Figure 1.   p21 immunostaining in normal bronchial glands. Magnification: ×500.

Immunostaining of neoplastic tissues for p21 was considered positive (more than 1% of cells) in 48 specimens (80%). Immunoreactivity for p21 was always detected in a few isolated cells in the remaining specimens. The percentage of immunoreactive cells varied from one specimen to another, ranging from 1% to 60% (Table 1). Immunoreactivity was greatest in well-differentiated areas of squamous carcinomas (Figure 2) and in papillary areas of adenocarcinomas. A few reactive cells were detected in the intratumoral lymphoid infiltrate. In 44 (73.3%) specimens with more than 5% of positive cells (as also previously reported [28]), p21 was considered to be overexpressed (Table 2). Two (3.3%) of the 44 p21-overexpressing specimens did not overexpress p53. In addition, both p21 and p53 proteins were undetectable in seven (11.7%) of the 60 investigated specimens (Table ). Loss of p21 occurred more often (22.7%) in p53-positive than in p53-negative (18.4%) tumors (Table ).

View larger version (114K): [in this window] [in a new window]   Figure 2.   A representative specimen of lung squamous carcinoma showing p21 immunostaining. An area of well-differentiated tumor with positive staining for p21 and an adjacent area of poorly differentiated tumor with no staining for p21 are visible. Magnification: ×500.

No relationship was found between p21 overexpression and tumor stage or TNM status, but histologic grading was slightly correlated with p21 status (P = 0.51). No significant difference was noted between squamous carcinomas and adenocarcinomas. We further divided our specimens into two groups. The first group included all specimens with p21 immunostaining detected in more than 5% of cells (p21+). The second group contained the remaining specimens (p21). Analysis of the data using such arbitrary cutoff values was highly statistically significant in that p21 status correlated statistically with short-term survival (P < 0.001) (Figure 3). The mean survival time for p21+ patients was 36 mo, as compared with 14 mo for p21 patients. The 5-yr survival rate for p21 patients was 10%, as compared with 38% for p21+ patients (Table 3). No significant correlation was found between tumor grade and the overall survival of patients. All patients who overexpressed p53 with associated loss of p21 were dead within 1 yr after undergoing surgery.

View larger version (19K): [in this window] [in a new window]   Figure 3.   Survival percentage curves for lung-cancer patients, based on different p21 expression levels and constructed through Kaplan-Meier analysis. Statistical significance of these data was measured with analysis of variance and the chi-square technique, generating a value of P < 0.001.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

p21 is generally considered to be a downstream target of p53. However, it has been shown that p21 can be activated independently of the expression of p53. Expression of p21 can be induced in p53-null cells by different stimuli such as serum, growth factors, drugs, and growth-arrest inductors (22).

In this study, we found p21 protein expressed both in normal and in neoplastic tissues. p21 immunoreactivity was detectable in a low percentage of well-differentiated cells such as ciliated cells and Type II alveolar cells in normal bronchial tissues. However, we found p21 nuclear staining in 80% of the tumors we investigated. The majority of the investigated specimens (73.3%) met the established criteria (number of positive cells > 5%) for overexpressing status.

We chose immunohistochemistry for this investigation because it provided the advantage of permitting visualization of the staining pattern of each section without contamination by nonneoplastic elements such as stromal, endothelial, and inflammatory cells, which always affects the nucleic acid-based approach. Other investigators, using an immunohistochemical and RNA-based double approach, showed similar results and demonstrated that p21 protein and RNA expression are strictly correlated (29).

Our results showed significant differences in mean (P < 0.001) and overall (P < 0.001) survival time of patients of different p21 status. This finding is inconsistent with the report of DiGiuseppe and colleagues (30), who described the p53-independent expression of p21 in a series of pancreatic carcinomas, but is consistent with the report of El-Deiry and coworkers (31), who described low to absent levels of the transcript of p21 in four colorectal carcinomas with mutated and overexpressed p53.

In addition, we found that p21 immunostaining was detectable mostly in well differentiated areas of the tumors we investigated. However, the correlation between p21 expression and tumor grade, even if statistically significant, was very weak (P = 0.51).

Our results suggest a correlation between p21 expression and the degree of malignancy of NSCLC. This suggestion is strengthened by the significant differences found in the 5-yr survival rate of our patients and in the mean and overall survival times of patients of different p21 status (Table ). In addition, even if p21 overexpression correlated slightly with tumor grade, no significant correlation was found between tumor grade and the overall survival rate of the patients. When we considered the p21 status of our p53-overexpressing patients, we found that all p53+ p21 patients were dead within 1 yr of surgical resection. This last datum, even if the small number of p53+ p21 specimens does not allow any statistical consideration, suggests that p21+ status may be a favorable prognostic marker for patients with altered p53 expression.

To the best of our knowledge, this is the first report of a study that has followed the relationship between p21 expression and NSCLC. In conclusion, our results suggest that p21 may play a predictive role in the clinical behavior of NSCLC, independent of p53 status and the histologic grade of the neoplasm. However, additional studies with a larger number of patients, including patients with small-cell lung cancers, are needed to confirm these observations.