Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Cellular proliferation is regulated by protein complexes composed of cyclins and cyclin-dependent kinases (cdks). These complexes are enzymatically active phosphorylating cell-cycle regulatory elements, such as the retinoblastoma proteins. In these interactions, cdks act as catalytic subunits and cyclins play the role of regulatory molecules, peaking in specific phases of the cell cycle (1). At present, five major families of cyclins (termed A, B, C, D, and E) have been isolated and characterized. Cyclins A and B1 and 2 expression increase later in the cell cycle, during the S and G2 phases, and are considered regulators of the transition to mitosis. On the other hand, cyclins C, D1-3, and E reach their peak of synthesis and activity during the G1 phase, and are believed to regulate the G1-to-S phase transition (1). Because this is an important restriction point for mammalian cell growth control, these cyclins are considered to be potential oncogenes.

Of interest, the cyclin D1 gene is mapped in one of the most frequently amplified chromosomal regions (11q13) in human carcinomas (2), and recent research shows that the cyclin D1 gene is frequently amplified in breast and lung carcinomas (3). In addition, a frequent detection of cyclin D1 overexpression in the absence of apparent genetic amplification in breast cancers has been reported (6). This overexpression is known to be associated with advanced local invasion and the presence of lymph node metastases in laryngeal, head, and neck carcinomas (7).

Some authors studying ectopically expressed cyclin D1 in cell cultures have suggested that cyclin D1 alone is required for protein Retinoblastoma (pRb) inactivation and G1 progression (8, 9) because it has demonstrated its ability to immortalize cells in vitro (10, 11). In the G1 phase, cyclin D1 is present in a quaternary complex with cdk4, a small protein of 21 kD that acts as a cdk4 inhibitor, and with the proliferating cell nuclear antigen (PCNA), which is involved in the mechanism of DNA excision repair (12). Thus, cyclin D1 may modulate the activities of PCNA.

Lung cancer is now associated with one of the worst prognoses among human malignancies. Approximately 30% of patients are candidates for radical lung resection, and the outcomes for these patients often appear to be unpredictable. New approaches to the management of this disease are required to better define the profile of patients eligible to undergo curative surgical resection.

To study the potential prognostic role of cyclin D1 in lung cancer, we developed a polyclonal antibody raised against cyclin D1 to evaluate its expression in a group of 135 patients with lung cancer who had surgical resection. In addition, we compared cyclin D1 expression with the expression of PCNA, which is considered a reliable index of the proliferation rate and which interacts with cyclin D1 in the G1 phase.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Lung Cancer Specimens

A total of 135 formalin-fixed, paraffin-embedded lung-cancer specimens were obtained from patients who underwent a surgical resection (lobectomy or pneumonectomy) in the Department of Thoracic Surgery of the V. Monaldi Hospital of the University of Naples (Italy) or in the Department of Cardio-Thoracic Surgery of the University Hospital of Vienna (Austria) between 1989 and 1993. Fresh-frozen tissues were also available from 54 of the 135 patients. All specimens available during the time frame (1989-1993) from patients who had not received chemotherapy or radiotherapy prior to surgical resection were studied. Follow-up data were collected from the Central Institute of Statistics of Austria (University of Vienna, Vienna), hospital charts, and periodic interviews with patients and their families.

The histologic diagnoses and classifications of the tumors were based on World Health Organization criteria (13), and the postsurgical pathologic TNM stage was determined using the guidelines of the American Joint Committee on Cancer (14).

Antibody

A rabbit polyclonal immune serum against cyclin D1 was produced by immunizing rabbits with a bacterially expressed glutathione S-transferase-full-length cyclin D1 fusion protein. Expression of the fusion protein was performed as previously reported (15, 16).

Immunoprecipitation

Immunoprecipitation was carried out as previously described (17).

In Vitro Transcription Translation

In vitro transcription of the cyclins D1, D2, and D3 complementary DNA (cDNA) clones was performed by a T7 RNA polymerase capping reaction. Briefly, after phenol/ chloroform extraction and ethanol precipitation, the transcription products were used as substrates for in vitro translation using a rabbit reticulocyte lysate (Promega, Madison, WI) and ³⁵S-methionine as a radioactive label.

Western Blot Analysis

The amount of 1 g of each frozen lung cancer tissue sample was sectioned and quickly homogenized at 4°C in 250 mM NaCl, 50 mM Tris (pH 7.4), 5 mM ethylenediamenetetraacetic acid, 0.1% (vol/vol), Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 50 mM NaF, 0.5 mM Na₃VO4, 10 mg/ml leupeptin, and 50 mg/ml aprotinin. The homogenates were cleared by centrifugation for 15 min at 13,000 × g at 4°C, and the total protein in the extracts was determined. Western blot analysis was performed as previously described (18).

Immunohistochemistry

Briefly, sections from each specimen were cut at 3 to 5 µm, mounted on glass, and dried overnight at 37°C. All sections were then 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. Tissue sections for cyclin D1 immunostaining were heated sequentially twice in a microwave oven for 5 min each at 700 W in citrate buffer (pH 6) and then blocked with diluted 10% normal goat antirabbit serum (Vector Laboratories, Burlingame, CA). Slides were incubated for 1 h at room temperature with the rabbit polyclonal immune serums raised against cyclin D1 at a 1:100 dilution, and then were incubated with diluted goat antirabbit biotinylated antibody (Vector Laboratories) for 30 min at room temperature.

Sections for PCNA detection were quenched in 0.5% hydrogen peroxide and blocked with diluted 10% normal horse antimouse serum (Vector Laboratories). Monoclonal mouse anti-PCNA antibody PC10 (dilution 1:50) (DAKO Corp., Carpinteria, CA) was used. Incubation time was 60 min at room temperature. After washing in PBS, slides were incubated with diluted horse antimouse biotinylated antibody (Vector Laboratories) for 30 min at room temperature.

All of the slides were then processed by the ABC method (Vector Laboratories) for 30 min at room temperature. Diaminobenzidine was used as the final chromogen, and hematoxylin was used as the nuclear counterstain. Negative controls for each tissue section were made by substituting the primary antibody with the respective preimmune serum for cyclin D1ab or by leaving out the primary antibody for PC10. All samples were processed under the same conditions. Three pathologists evaluated the staining pattern of the proteins and scored each specimen for the percentage of positive nuclei: score 0, undetectable level; score 1, 1 to 30% of positive cells (low expression); score 2, 30 to 60% of positive cells (medium expression); and score 3, more than 60% of positive cells (high expression). A tumor was considered cyclin D1-negative if there was no nuclear staining in any neoplastic cell, regardless of cytoplasmatic staining, while admixed nonneoplastic elements did show nuclear immunoreactivity. If the latter were negative as well, the stain was considered uninterpretable. A cutoff of 1% of positive cells was also adopted. Specimens with less than 1% of positive cells were then included in the first group (score 0, undetectable expression). Analysis of the data by means of such necessarily arbitrary cutoffs was highly statistically significant. At least 20 high-power fields were chosen randomly, and 2,000 cells were counted.

Statistical Analysis

We performed a statistical analysis to investigate the relationship between clinico-pathologic parameters (age, gender, histotype, TNM status, tumor stage, histologic grading, and postoperative radiotherapy and/or chemotherapy), cyclin D1 expression, PCNA expression and patient survival time. Kaplan-Meier survival curves were constructed, and survival distributions were compared by means of the log-rank test. Linear-by-linear and Kruskal-Wallis association tests were used to assess any possible correlation between clinico-pathologic parameters, PCNA expression, Western blot cyclin D1 expression levels, and immunohistochemical data.

Statistical significance was declared if the P value was < 0.05. The analysis was performed with Stata 5.0 software (Stata Statistical Software; Stata Corp., College Station, TX).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Characterization of the Antibody Raised against Cyclin D1

The specificity of this antibody was confirmed by an immunoprecipitation experiment. As Figure 1 shows, using cell lysates from ³⁵S-labeled ML-1 cells, cyclin D1 migrated with a molecular weight of ~ 36 kD. Immunoprecipitation performed using the corresponding prebleeding normal rabbit serum was used as a negative control. Lack of cross-reaction of cyclin D1 immune serum with other members of the cyclin D family was confirmed by immunoprecipitation of this antibody with the in vitro translated forms of the cDNAs coding for cyclins D1, D2, and D3. Our serum was able to immunoprecipitate only the in vitro translated form corresponding to cyclin D1 (Figure 2).

View larger version (51K): [in this window] [in a new window]   Figure 1.   Sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis (12%) of cellular extracts from 35S- labeled ML-1 cells after immunoprecipitation of cyclin D1 protein by means of the rabbit polyclonal (anti-cyclin D1) immune serum. Normal rabbit serum (NRS) was used as negative control.

View larger version (48K): [in this window] [in a new window]   Figure 2.   In vitro translated (IVT) forms of cyclins D1, D2, and D3 (lanes 1, 3, and 5) were immunoprecipitated with the polyclonal serum against cyclin D1 (lanes 2, 4, and 6). Only the polyclonal antibody was able to immunoprecipitate the in vitro translated form corresponding with cyclin D1.

Cyclin D1 and PCNA Expression in Lung Cancer

A total of 135 patients was included in the study. Most patients were male (78.5%), and the mean age was 62 years. A total of 23 patients underwent postoperative chemotherapy and the same number of patients underwent postoperative radiotherapy. Review of the clinico-pathologic reports showed that 58 tumors were adenocarcinomas (43%), 54 were squamous carcinomas (40%), 22 were small-cell carcinomas (16.3%), and the histologic type of one specimen was unclear. A total of 43 tumors (31.9%) were classified as stage I, 40 tumors (29.6%) as stage II, 28 tumors (20.7%) as stage IIIa, 7 tumors (5.2%) as stage IIIb, and 8 tumors (5.9%) as stage IV. TNM status was incomplete in nine patients who therefore were not included in the stage groupings. In addition, 52 tumors were graded low differentiated (38.5%), 34 tumors medium differentiated (25.2%), and 5 tumors well differentiated (3.7%). Histologic grading was not available for the remaining specimens.

Cyclin D1 immunoreactivity was localized mostly in nuclei, but occasional cytoplasmatic spread was observed. The level of concordance expressed as a percentage of agreement between the observers was 94.8% (128 specimens of 135). In the remaining specimens, there was agreement between two of the three observers. In these specimens, the score was obtained from the opinions of the two agreeing investigators. Cyclin D1 immunostaining was considered interpretable in 98 specimens according to the criteria described in MATERIAL AND METHODS. In all of these specimens at least some non-neoplastic elements were clearly stained, but the percentage of positive non-neoplastic cells was relatively low compared with neoplastic elements. Sections for PCNA immunostaining were available in 95 of 98 specimens. A relatively small percentage of the 98 specimens exhibited undetectable expression levels of cyclin D1 (12 specimens, 12.2%). Low expression levels of cyclin D1 were recognized in 36 specimens (36.7%) and medium expression levels were detected in 24 specimens (24.5%), whereas 26 specimens (26.5%) exhibited a high expression level (Figure 3A). PCNA expression was at a detectable level in all 95 specimens studied, with nuclear expression levels varying from low (18 specimens, 18.9%), to medium (39 specimens, 41.1%), to high expression levels (38 specimens, 40%) (Figure 3B).

View larger version (134K): [in this window] [in a new window]   View larger version (124K): [in this window] [in a new window]   Figure 3.   (A) Cyclin D1 representative immunostaining in a lung squamous carcinoma (×300). (B) PCNA immunostaining in a lung adenocarcinoma (×300).

A statistically significant correlation was found between the expression of cyclin D1 and PCNA (P = 0.0004).

Western blot analysis performed on 54 specimens showed different expression levels of cyclin D1, ranging from undetectable levels (7 specimens, 13%) to high expression levels (18 specimens, 33.3%) (Figure 4). These results correlated with the immunohistochemical findings (P = 0.0003). No correlation was found between the immunohistochemical expression of either cyclin D1 or PCNA and any clinico-pathologic factor.

View larger version (13K): [in this window] [in a new window]   Figure 4.   Western blot analysis showing different cyclin D1 expression levels.

When we analyzed our data statistically according to clinico-pathologic factors, cyclin D1 and PCNA status, and survival distributions, the percentage of cells positive for cyclin D1but no other variablewas found to correlate with patient survival times (P < 0.00005). High cyclin D1 expression levels were associated with the poorest outcome for the patients (Figure 5).

View larger version (15K): [in this window] [in a new window]   Figure 5.   Kaplan-Meyer survival percentage curves for lung cancer patients, constructed according to different cyclin D1 expression status.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Several studies have proposed that cyclins could act as oncogenes by forcing cells to evade the normal growth control (19). Overexpression or amplification of cyclin D1 has been demonstrated in several cancers, including B-cell leukemia; lymphoma; and esophageal, colorectal, breast, and lung carcinomas (20).

We developed an immunohistochemical assay specific for cyclin D1 that is suitable for formalin-fixed and paraffin-embedded sections. Western blot analyses confirmed the specificity of this assay by correlating with the immunohistochemical results in a highly significant manner (P = 0.0003).

We found cyclin D1 to be expressed in more than 60% of the neoplastic cells in 26.5% of our specimens. A total of 24.5% of the specimens still showed cyclin D1 expression in a percentage of cells ranging from 30 to 60%, whereas 36.7% of the specimens expressed cyclin D1 in less than 30% of the cells and 12.2% of the specimens did not express this protein in more than 1% of the evaluated cells. In addition, we found a direct relationship between cyclin D1 and PCNA immunodetection (P = 0.0004) that correlated cyclin D1 overexpression with a higher tumor-proliferation rate. These results confirm the suggestion of Shauer and colleagues, who hypothesized the possibility that lung tumors may evade cell-cycle controls by abnormal expression of cyclin D1 (5). Others failed to demonstrate any relationship between cyclin D1 expression and the proliferation rate of the tumor, as assessed by ki 67 immunostaining (27).

When we analyzed our data statistically, cyclin D1 expression was found to be a negative prognostic marker (P < 0.00005) whose expression correlates with a shorter patient survival time. Betticher and associates reported that cyclin D1 expression correlates with the grade of differentiation of the tumor, with the amount of lymphocytic infiltration, and with a reduction in local relapse data, but not with patient outcomes (26). Other recent studies on lung-cancer patients have reported contrasting results, either correlating cyclin D1 expression with a shorter survival time (28) or stating the usefulness of cyclin D1 as a prognostic marker (27).

Our finding of a correlation between cyclin D1 expression and a higher proliferation rate and reduced survival time has been confirmed indirectly in several studies that previously demonstrated the oncogenic potential of cyclin D1. It has been shown that constitutively overexpressed cyclin D1 can shorten the G1 phase in rodent cells (29). On the other hand, microinjection of cyclin D1 antibodies or antisense plasmid into dividing cells blocks them in the G1 phase (29, 30). In addition, it has been shown that cyclin D1 induction is sufficient to complete a cell-cycle shortening of the G1-S phase in breast cancer cells arrested in G1 phase.

This study strengthens the potential oncogenic role of cyclin D1 expression in lung carcinogenesis. However, its role in the prognosis for lung cancer patients continues to be unclear because of the conflicting results reported in other recent studies (26). Additional study of larger numbers of patients is required to confirm the suggestions offered by the present work.