INTRODUCTION

Since its initial discovery by Waisserman and Taylor in 1966, calbindin has been reported in a wide variety of species and tissues, including cerebellum, kidney, placenta, thymus gland, and neuroendocrine cells of the gastrointestinal tract, lung, pancreas, and thyroid gland (1, 2, 3, 4, 5, 6). Calbindin belongs to a family of intracellular proteins that have affinity for calcium. Other family members include S-100 protein, calmodulin, parvalbumin, and troponin C (7). Calbindins are acidic proteins that are heat stable. Two major subclasses have been described: D9k and D28k. Calbindin D28k is encoded by a gene located on chromosome 8 that is highly conserved in evolution (7). Ca-D28k is a single polypeptide chain that consists of 261 amino acid residues, with four high-affinity calcium-binding sites.

The physiologic function of Ca-D28k is unknown, although it has been postulated that Ca-D28k (and other calbindins) constitutes an intracellular buffering mechanism to adjust calcium to a physiologic level (7). Because calcium is thought to play an important role in apoptosis by stimulating calcium/magnesium-dependent endonucleases and by modulating calcium/calmodulin-dependent enzymatic activity (8), Ca-D28k may play a role in modulating apoptosis. Dowd and colleagues (8), using a thymoma cell line, showed that Ca-D28k overexpression decreases sensitivity to apoptosis, presumably by interfering with calcium fluxes that induce programmed cell death.

Previous studies of Ca-D28k expression in lung cancer have been few and limited in scope (9), and the prognostic relevance of Ca-D28k expression in lung cancer has not been examined. The purpose of this study was to assess the prevalence of Ca-D28k expression in lung carcinomas of various histologic types and to determine whether Ca-D28k expression has any prognostic significance.

MATERIALS AND METHODS

Paraffin blocks of formalin-fixed biopsy and surgical resection tissues were obtained from the archives of The Methodist Hospital, Houston, Texas. A representative paraffin block of each neoplasm was selected. The histologic classification of each tumor was based on the World Health Organization criteria (10). Stage was determined using the TNM system (11).

Medical charts and cancer registry data for each patient were reviewed for age, sex, smoking history, clinical and pathologic staging, and survival. Survival was correlated with age, sex, histologic pattern of carcinoma, stage, and presence or absence of Ca-D28k immunostaining. The minimum follow-up period for each surviving patient was at least 5 years. Patients who died within 3 months after surgical excision or who died from causes other than lung cancer within 5 years after surgery were excluded from the final statistical analysis.

Immunohistochemical staining for Ca-D28k was done using routinely processed, paraffin-embedded tissue sections. Briefly, after slides were treated with hydrogen peroxide to block endogenous peroxidase activity, they were washed in distilled water and immersed in 0.01 m sodium citrate buffer (pH 6.0) for 20 min at 95° C, followed by rinsing in distilled water and phosphate-buffered saline. The slides were then incubated for 2 h with a monoclonal antibody specific for Ca-D28k (dilution 1:250, code no. 300; Swant Swiss Antibodies, Bellinzon, Switzerland). Staining for Ca-D28k was carried out with an avidin-biotinylated peroxidase method according to the recommendations of the manufacturer (Vector Laboratories, Burlingame, CA). A lung carcinoma was considered to be positive for Ca-D28k if at least 10% of the tumor cells exhibited nuclear staining. The number of positive tumor cells was graded semiquantitatively using the following five groups: 0, less than 10%, more than 10 to 25%, more than 25 to 50%, more than 50 to 75%, and more than 75%. All stained cells were recorded as positive regardless of staining intensity.

Overall patient survival was calculated from the date of operation by Kaplan-Meier cumulative survival plot using SPSS 7.5 software (Chicago, IL). Overall survival was compared across levels of prognostic factors using the log-rank test. All statistical tests were two sided.

RESULTS

A total of 452 patients with primary lung carcinoma diagnosed between 1984 and 1991 were studied. The demographic characteristics, smoking history, and survival data are shown in Table 1. A total of 425 patients (94%) had a positive history of smoking. A total of 151 patients (33%) were alive 5 years after diagnosis, with a mean follow-up period of 44 months.

TABLE 1 Histologic Type and Demographic Characteristics of 452 Patients with Lung Cancer
Full size table

The lung carcinomas assessed in this study were as follows: 178 (39%) adenocarcinoma, 113 (25%) squamous cell, 73 (16%) large cell undifferentiated, 62 (14%) bronchioloalveolar, 17 (4%) small cell, and 9 (2%) adenosquamous. The distribution of stage, treatment and mean survival is shown in Table 2. The mean survival was best for patients with bronchioloalveolar carcinoma (63 months) and worst for patients with small cell carcinoma (29.8 months).

TABLE 2 Histologic Type, Pathologic Stage, Treatment, and Survival of 452 Patients with Lung Cancer
Full size table

Table 3 summarizes the frequency of Ca-D28k expression for the different types of lung carcinoma. A total of 335 of the 452 (74%) lung carcinomas were positive for Ca-D28k (Fig. 1). There was no difference in Ca-D28k immunostaining between biopsy and resection specimens. There was no significant correlation between the prevalence of Ca-D28k expression and histologic type of carcinoma. The prevalence of immunohistochemically detectable Ca-D28k in each tumor type was as follows: 8 (88%) adenosquamous carcinoma, 48 (77%) bronchioloalveolar carcinoma, 86 (76%) squamous cell, 13 (76%) small cell carcinoma, 135 (75%) adenocarcinoma, and 45 (61%) large cell undifferentiated carcinoma.

TABLE 3 Histologic Type of Lung Cancer, Ca-D28k Expression, and Patient Survival
Full size table
FIGURE 1
figure 1

Nuclear and cytoplasmic expression of Ca-D28k. A, squamous cell carcinoma; B, adenocarcinoma; C, bronchioloalveolar carcinoma; D, small cell carcinoma; E, normal bronchial epithelium.

Full size image

The mean survival for Ca-D28k–positive tumors (49 months) was slightly better than for Ca-D28k–negative tumors (45 months), with a value approaching statistical significance (Table 3; P = .061). Ca-D28k expression significantly correlated with survival for patients with adenocarcinoma. Patients with Ca-D28k–positive tumors (mean, 44.1 months) had improved survival over patients with Ca-D28k–negative tumors (mean, 29.2 months; P = .036) Table 3 and Fig. 2. This was also true for patients with Stages I and II adenocarcinoma (P = .033), but not for patients with Stages III and IV adenocarcinoma. There were no other statistically significant differences in patient survival between Ca-D28k–positive and Ca-D28k–negative lung tumors of other histologic types.

FIGURE 2
figure 2

Kaplan-Meier survival curves for patients with lung adenocarcinoma.

Full size image

Ca-D28k expression was analyzed in relation to various clinical characteristics including sex, smoking history, TNM stage, and therapy (Table 4). None of these characteristics correlated with Ca-D28k expression.

TABLE 4 Demographic Characteristics of 452 Patients with Lung Cancer Positive or Negative for Ca-D28k
Full size table

DISCUSSION

Previous studies of Ca-D28k expression assessed in lung cancer have been limited in scope (9). In this study, we assessed Ca-D28k expression in a large number of lung carcinomas of various histologic types and have compared Ca-D28k expression with patient demographic characteristics and survival.

We evaluated Ca-D28k expression using an immunohistochemical method and detected Ca-D28k expression in 335 of 452 (74%) tumors. The prevalence of Ca-D28k expression was similar for all histologic types of lung carcinoma assessed: adenosquamous (88%), bronchioloalveolar (77%), squamous cell (76%), small cell (76%), adenocarcinoma (75%), and large cell undifferentiated (61%). These results are in agreement with those of Watanabe and colleagues (9), who measured the concentration of Ca-D28k in lung cancer cell lines using enzyme immunoassay methods. They reported Ca-D28k expression in all lung cancer cell lines; however, the prevalence of Ca-D28k expression was highest in small cell carcinoma. A possible explanation for the discrepant results between our study and Watanabe et al.’s (9) is the use of different detection methods or the different types of specimens assessed.

Our findings contrast with those of Katsetos and colleagues (12), who studied immunohistochemical expression of Ca-D28k in a small study of 41 lung tumors. In that series, Ca-D28k expression was observed only in neuroendocrine tumors, including four small cell carcinomas, two atypical carcinoids, and two typical carcinoids assessed. Katsetos et al. (12) did not demonstrate Ca-D28k in other lung carcinoma types. Other reports also have found that Ca-D28k is expressed in normal endocrine cells of stomach (4, 5), respiratory epithelium (13), pancreas (14), duodenum (15), and neuroendocrine tumors of the gastrointestinal tract (12). Thus, others have suggested that Ca-D28k may be a useful marker of neuroendocrine differentiation (12). In contrast, we observed that CaD28k is present in a wide variety of lung carcinomas of all histologic types. Our results demonstrate that Ca-D28k expression is not limited to neuroendocrine tumors and may have more widespread functional significance.

We analyzed the prognosis of patients with lung carcinoma and Ca-D28k expression. Patients with Ca-D28k–positive adenocarcinomas had significantly better survival (mean, 44.1 months) than patients Ca-D28k–negative adenocarcinomas (mean, 29.2 months; P = .036). The difference in survival was also significant for patients with localized (Stages I and II) adenocarcinomas (P = .033). Ca-D28k expression did not significantly correlate with patient survival for disseminated adenocarcinomas or other histologic types of lung carcinoma. In some ways, this result is counterintuitive. If Ca-D28k overexpression is thought to play a role in inhibiting apoptosis (8), one might suspect that Ca-D28k–positive tumors would have a worse prognosis. However, other factors are involved that might be more important in determining patient prognosis, for example, the primary treatment modality (surgical excision alone versus surgical excision and adjuvant therapy) and the adequacy of surgical excision margins.

The true functional role of Ca-D28k is not known. In addition to its role in inhibiting apoptosis, some observations suggest that Ca-D28k may be involved in the regulation of the secretory process (1, 7). Adenocarcinomas with preservation of the mucin secretion are better differentiated and have a better prognosis than poorly differentiated adenocarcinomas. It is therefore possible that Ca-D28k–positive adenocarcinomas represent a group of tumors that are better differentiated than Ca-D28k–negative adenocarcinomas, although we did not observe a clear correlation between Ca-D28k expression and differentiation histologically in this study.

In prior studies, conflicting results regarding the localization of Ca-D28k at the cellular level have been published. In some studies of other malignant neoplasms and normal tissues, Ca-D28k expression was thought to be restricted to the cell membrane (2, 3, 4, 5, 6). In contrast, others (12, 13, 14) have reported results similar to our own. We observed that Ca-D28k is expressed within the nuclei and cytoplasm of lung carcinoma cells of all histologic types and in normal bronchial epithelium. The difference in the distribution of Ca-D28k within the cell is most likely a manifestation of different roles that the protein has within cells.