Lung cancer is a malignant tumor characterized by high morbidity and mortality rates, accounting for ~1.2 million mortalities annually worldwide, which usually result from disease recurrence and metastasis (1). Although diagnostic methods and treatments have markedly improved in recent years, the 5- and 10-year survival rates remain at <15 and <7%, respectively (2). At present, the lack of appropriate molecular diagnostic tools to predict the potential metastasis of lung cancer represents a major clinical obstacle. Therefore, identification of biomarkers that accurately predict future recurrence and metastasis of lung cancer may improve treatment strategies.

Carboxypeptidase E (CPE) is a metal ion-dependent exopeptidase that is predominantly expressed in endocrine and nervous tissues, which converts the prohormones secreted by endocrine or nerve cells, such as adrenocorticotropin/lipotropin (ACTH/LPH), proinsulin, opiomelanocortin and enkephalin, into an active form or into neuropeptides (3–5). Recent studies have demonstrated that abnormal expression of CPE occurs in epithelium-derived cancer tissues, including liver cancer, renal clear cell carcinoma, colorectal cancer, cervical cancer and melanoma (6–14). In 2011, a novel form of CPE, N-terminal truncated carboxypeptidase E (CPEΔN), was identified in hepatocellular carcinoma (HCC) (15,16). Truncated CPE interacts with histone deacetylase (HDAC) 1 and HDAC2 to form a complex, which regulates the expression of metastasis-associated proteins. Furthermore, it was reported that CPEΔN was an independent predictor for recurrence and metastasis of HCC. HCC patients with high CPEΔN expression levels exhibited significantly higher recurrence rates in the 2 years following surgery and lower median survival times than patients with low CPEΔN expression levels (15).

Similar findings have been observed in primary pheochromocytomas/paragangliomas (PHEO/PGL) and colorectal cancer patients (17,18). Previous studies, which have used quantitative polymerase chain reaction to analyze CPEΔN expression, revealed that elevated CPEΔN expression is a statistically significant predictor of poor prognosis (17,18). Since HCC, PHEO/PGL and colorectal cancer are extremely different tumors with distinctive tumor origins, these findings suggest that CPEΔN may be a predictor of metastasis with a broad spectrum.

To evaluate the function of CPEΔN in lung adenocarcinoma, in the present study, CPEΔN expression was analyzed in lung adenocarcinoma tumors by western blot analysis and immunohistochemistry. It has been demonstrated that CPEΔN expression was associated with lymph node metastasis and distant metastasis of lung adenocarcinoma, and that the three-year tumor-free survival rates were significantly lower in patients with CPEΔN expression than in those without CPEΔN expression. The present study aimed to evaluate the potential of CPEΔN as a biomarker for predicting future metastasis of lung adenocarcinoma.

CPE is a multifunctional protein that exhibits both enzymatic and non-enzymatic functions. CPE exists in three different forms, which are expressed in various subcellular localizations with distinct functions (18). The first type is released by secretory granules, which results in the cleavage of basic residues to generate mature peptide hormones and neuropeptides. Recently, Skalka et al (22) demonstrated that this secreted form of CPE forms a complex with the Wnt3a ligand and frizzled receptor to negatively regulate the canonical Wnt signaling pathway (22). The second type of CPE exists in the membrane, anchored at the trans-Golgi network, which functions as a sorting receptor for prohormones (23). The third type of CPE is located in both the cytoplasm and nucleus, and is involved in cell signal transduction and transcription regulation. The function of this type of CPE is the most diverse, however, further investigation is required with regard to this protein.

In the present study, the association between CPEΔN expression and prognosis of lung adenocarcinoma patients was evaluated. The results revealed that two forms of CPE, full-length CPE and CPEΔN, are co-expressed in the lung adenocarcinoma tissues. CPEΔN expression was identified in the primary tumor tissues of 62.1% (59/95) patients and non-tumor tissues of 5.26% (5/95) patients. Furthermore, multivariate Cox regression analysis demonstrated that CPEΔN expression was closely associated with the occurrence of lymph node and distant metastasis. Patients with positive CPEΔN expression exhibited significantly lower disease-free survival rates than patients with negative CPEΔN expression (87.5 vs. 69.2%, respectively; P=0.009). These findings suggested that CPEΔN expression indicates a higher risk of recurrence and metastasis in primary lung adenocarcinoma.

Western blot analysis requires a large amount of sample tissue, which is a major problem in clinical research. Development of simplified methods for the analysis of CPEΔN or CPE expression is required. In our previous studies (Sun et al, unpublished data), the subcellular localization of CPEΔN was investigated using confocal microscopy, which revealed that CPEΔN was localized to the nucleus in H1299, A549, 95D, H1395 and Calu3 lung cancer cell lines, whereas full-length CPE was predominantly localized in the cytoplasm. Therefore, it was hypothesized that similar findings may be observed in primary lung adenocarcinoma. The hypothesis was tested with tissue microarray. Since one array can simultaneously test 75 samples along with corresponding pericarcinoma tissues, 75 patients with intact overall survival and progression-free survival profiles were selected. Patients were divided into two groups, a nuclear CPE expression group and a cytoplasmic CPE expression group. A difference in overall survival rate was identified between the nuclear and cytoplasmic CPE expression groups (46.3 vs. 64.7%, respectively); however, this was not statistically significant (P=0.125). However, future studies that include more lung adenocarcinoma patients and samples are required to validate the results of the present study.

In conclusion, the results of the present study demonstrated that CPEΔN expression is associated with poor prognosis in lung adenocarcinoma. These findings may improve understanding with regard to the underlying molecular mechanisms of CPE/CPEΔN expression, which promote lung cancer metastasis. Thus, the evaluation of CPEΔN expression status may aid to identify primary lung adenocarcinoma patients that require more intensive treatment. Furthermore, CPEΔN may present a potential target for therapeutic intervention in the future.