Clinical implication and prognostic significance of the tumor suppressor TSLC1 gene detected in adenocarcinoma of the lung

Authors

  • Kazuya Uchino M.D,

    1. Department of Cardiothoracic Surgery, Kobe University Medical School, Kobe, Hyogo, Japan
    2. Department of Thoracic Surgery, Hyogo Medical Center for Adults, Akashi, Hyogo, Japan
    Search for more papers by this author
  • Akihiko Ito M.D., Ph.D,

    1. Department of Pathology, Osaka University Medical School/Graduate School of Frontier Bioscience, Suita, Osaka, Japan
    Search for more papers by this author
  • Tomohiko Wakayama M.D., Ph.D.,

    1. Department of Histology and Embryology, Kanazawa University Graduate School of Medical Science, Kanazawa, Ishikawa, Japan
    Search for more papers by this author
  • Yu-ichiro Koma Ph.D.,

    1. Department of Pathology, Osaka University Medical School/Graduate School of Frontier Bioscience, Suita, Osaka, Japan
    Search for more papers by this author
  • Tomoyo Okada M.D., Ph.D.,

    1. Department of Cellular Biochemistry and Biophysics Program, Memorial Sloan-Kettering Cancer Center, New York, New York
    Search for more papers by this author
  • Chiho Ohbayashi M.D., Ph.D.,

    1. Divisions of Surgical Pathology, Kobe University Medical School, Kobe, Hyogo, Japan
    Search for more papers by this author
  • Shoichi Iseki M.D., Ph.D.,

    1. Department of Histology and Embryology, Kanazawa University Graduate School of Medical Science, Kanazawa, Ishikawa, Japan
    Search for more papers by this author
  • Yukihiko Kitamura M.D., Ph.D.,

    1. Department of Pathology, Osaka University Medical School/Graduate School of Frontier Bioscience, Suita, Osaka, Japan
    Search for more papers by this author
  • Noriaki Tsubota M.D., Ph.D.,

    1. Department of Thoracic Surgery, Hyogo Medical Center for Adults, Akashi, Hyogo, Japan
    Search for more papers by this author
  • Yutaka Okita M.D., Ph.D.,

    1. Department of Cardiothoracic Surgery, Kobe University Medical School, Kobe, Hyogo, Japan
    Search for more papers by this author
  • Morihito Okada M.D., Ph.D.

    Corresponding author
    1. Department of Cardiothoracic Surgery, Kobe University Medical School, Kobe, Hyogo, Japan
    2. Department of Thoracic Surgery, Hyogo Medical Center for Adults, Akashi, Hyogo, Japan
    • Department of Thoracic Surgery, Hyogo Medical Center for Adults, Kitaohji-cho13-70, Akashi City, 673-8558, Hyogo, Japan
    Search for more papers by this author
    • Fax: (011) 81-78-929-2380


Abstract

BACKGROUND

Recently, the TSLC1 (tumor suppressor in lung cancer 1) gene has been identified as a novel tumor suppressor in human nonsmall cell lung carcinoma. To the authors' knowledge, the clinical relevance of TSLC1 gene expression has not been studied using patient data and surgical samples. The current study was designed to evaluate whether the TSLC1 gene can serve as a target for the prognostic determination of patients with pulmonary adenocarcinoma.

METHODS

A total of 38 patients who were surgically treated for proven primary lung adenocarcinoma were enrolled in the current study. Surgical specimens were examined for TSLC1 protein expression immunohistochemically and by Western blot analysis. The correlation between levels of TSLC1 expression and pathologic characteristics, as well as prognosis, was investigated.

RESULTS

All patients underwent a potentially curative resection of their tumor. TSLC1 antigen expression as evaluated by immunohistochemistry was confirmed by immunoblotting. The expression of TSLC1 protein was found to be inversely correlated with advanced disease stage, lymph node involvement, lymphatic permeation, and vascular invasion. The 4-year overall survival rates of patients with a tumor demonstrating high (> 70% positive cells [n = 14 patients]), intermediate (20–70% positive cells [n = 10 patients]), and low (< 20% positive cells [n = 14 patients]) expression of the TSLC1 antigen were 84%, 28%, and 7%, respectively. In addition, the disease-free survival of patients with a tumor that demonstrated a high percentage of TSLC1 protein-positive cells was reported to be significantly better than that of patients with a tumor that showed a low percentage of TSLC1 protein-positive cells.

CONCLUSIONS

The loss or reduction of TSLC1 expression in resected lung adenocarcinoma cases was associated with a poor prognosis, indicating that TSLC1 represents a central effector gene for controlling the biologic aggressiveness of the tumor and that it is an essential biomarker for predicting patient prognosis. These data may help to detect those patients at high risk for recurrence who might benefit from additional therapeutic strategies such as adjuvant therapy. Cancer 2003;98:1002–7. © 2003 American Cancer Society.

DOI 10.1002/cncr.11599

Lung carcinoma has recently become the most commonly reported cause of cancer death and its incidence is reported to be increasing throughout the world. Patients with nonsmall cell lung carcinoma represent approximately 80% of all lung carcinoma cases; it is these patients for whom generally radical surgery (RO resection) offers the only chance for cure. The identification of those patients who are at risk of recurrence after curative surgery or who are likely to benefit from supplementary treatment would be essential for planning cancer therapy. In addition to classic prognostic factors such as TNM staging, molecular markers are well worth identifying and investigating. Clarification of molecular targets for therapeutic approaches against nonsmall cell lung carcinoma is one of the more critical issues in attempting to improve the survival of these patients.

Mutations of the n-ras and k-ras-2 genes1 and alterations of the RB1, PPP2R1B, and p53 genes2–5 have been shown to be involved in human nonsmall cell lung carcinoma. In addition, loss of heterozygosity for chromosomes 3p, 11q and 13q has been demonstrated frequently in nonsmall cell lung carcinoma, implying that other tumor suppressor genes could be detected in these chromosomal regions.6–9 Loss of heterozygosity was observed at an especially high frequency at a chromosomal locus on 11q23. In 2001, the novel tumor suppressor gene TSLC1 (tumor suppressor in lung cancer 1) was identified on chromosome 11q23.2 by functional complementation of nonsmall cell lung carcinoma.10TSLC1, which encodes a member of the immunoglobulin (Ig) superfamily proteins and reveals significant homology with the neural cell adhesion molecule (NCAM)-1 and NCAM-2, is considered to mediate cell-cell interaction. Return of TSLC1 expression to the normal level in human lung adenocarcinoma cell lines was found to strongly suppress tumor formation and inactivation of TSLC1, although promoter methylation was noted in the majority of primary nonsmall cell lung carcinoma cases demonstrating loss of heterozygosity on 11q23.10

To our knowledge, there have been no reports to date regarding the role of the TSLC1 gene in the clinical setting of lung carcinoma. We therefore undertook the current study in patients with curatively resected adenocarcinoma, which is a major histologic subtype of lung carcinoma (the incidence of which recently has been reported to be escalating), to conclusively evaluate the prognostic importance of decreased TSLC1 expression. The major purpose of the current study was to investigate whether expression of the TSLC1 gene as a molecular marker could serve to identify adenocarcinoma patients at a high risk for failing to respond to standardized treatment.

MATERIALS AND METHODS

Tumor samples were obtained from 38 patients who underwent complete surgical resection of primary adenocarcinoma at Hyogo Medical Center for Adults. Patients who had received preoperative chemotherapy or radiotherapy were excluded. Preoperative evaluation included a detailed history and physical examination; biochemical profile; chest X-ray examination; bronchoscopy; and computed tomography (CT) scans of the chest, brain, and upper portion of the abdomen. Pathologic staging was determined according to the international TNM staging system.11 Intraoperative staging was performed by dissecting intrapulmonary, hilar, and mediastinal lymph nodes, and careful postoperative examination was performed by pathologists.12, 13 The histologic type of the tumor was determined by applying the World Health Organization classification. The status of lymphatic permeation by tumor and vascular involvement by tumor were abbreviated as the Ly factor and V factor, respectively. These factors were defined as follows: Ly(+) indicates positive lymphatic permeation, Ly(-) indicates negative lymphatic permeation, V(+): indicates positive vascular involvement, and V(-) indicates negative vascular involvement. From the representative sections stained with hematoxylin and eosin or with elastica van Gieson, the presence of tumor cells including emboli in endothelial-lined channels was determined by identification of intratumoral lymphatic permeation and vascular involvement. After discharge from the hospital, all patients were monitored at 3-month intervals for the first 2 years and at 6-month intervals thereafter. Follow-up assessment for signs of recurrence included physical examination, tumor marker, chest roentgenograms and CT scans. Patient follow-up was complete with regard to survival and recurrence in all patients.

Immunohistochemistry

Rabbits were immunized against the synthetic polypeptide containing 15 amino acids of the C-terminus of SgIGSF, a mouse orthologue of TSLC1.14 Four months later, the rabbit sera were purified with an affinity column containing the synthetic polypeptide and were used as an anti-TSLC1 antibody because the amino acid sequence of the C-terminus of SgIGSF was the same as that of TSLC1. The preparation and the sensitivity of the antibody were as described previously.15, 16

Samples were obtained both from the largest cut surface of each tumor and from normal lung tissues close to the tumor. After paraffin-embedded tumor specimens that had been fixed in neutral-buffered formalin were sectioned (4 μm thick), the sections were deparaffinized, rehydrated, and washed 3 times with phosphate-buffered saline (PBS) prior to every incubation procedure. The sections incubated in PBS containing 2% bovine serum albumin (PBS/2%BSA) for 30 minutes to prevent nonspecific immunoglobulin binding first were kept with primary antibodies (a rabbit anti-TSLC1 IgG [1:500 dilution] in PBS/2%BSA for 6 hours at 4 °C). Sites of primary antibody binding were visualized with biotinylated antirabbit IgG (1:100 dilution) (MBL, Nagoya, Japan) and peroxidase-conjugated streptavidin (1:400 dilution) (Dako, Kyoto, Japan), and developed in aminoethylcarbazole (AEC) (Dako). Nuclei were counterstained with hematoxylin. Negative immunohistochemical control procedures included: 1) omission of the primary antibodies and 2) replacement of the primary antibodies by normal rabbit or mouse IgG in appropriate concentrations. These control procedures gave negative results. Slides were evaluated by two independent pathologists with regard to the percentage of positively stained cells.

Western Blot Analysis

Protein samples were obtained from the largest cut surface of each tumor and from normal lung tissues adjacent to the tumor. After tissue homogenization in a buffer solution containing 50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 1% Triton X-100, and 1 mM phenylmethylsulfonyl fluoride, the extracted proteins were loaded into each lane of a sodium dodecyl sulfate (SDS)-polyacrylamide gel (10%). The gel then was electrophoresed and proteins were transferred to Immobilon (Millipore, Bedford, MA). Immunoblotting was performed using primary rabbit anti-TSLC1 antibody (1:500 dilution) in a buffer containing 20 mM Tris-HCl (pH 8.0), 150 mM NaCl, 0.05% Triton X-100, and 5% skim milk (Difco, Sparks, MD) (TBST/milk) for 3 hours at 4 °C. Secondary detection of primary antibody localization was accomplished using a horseradish-peroxidase–conjugated antirabbit whole IgG in TBST/milk for 2 hours at 4 °C, and then reacted with Western Lighting reagents (PerkinElmer Life Sciences, Boston, MA) before exposure. After stripping, the blots first were incubated with an anti-α-tubulin antibody (Sigma Chemical Company, St. Louis, MO) followed by a peroxidase-conjugated antimouse IgG antibody (MBL), and then were reacted with Western Lighting reagents (PerkinElmer) before exposure. Chemiluminescent intensities of specific signals were calculated with the FluorChem IS-8000 system (Alpha Innotech Corporation, San Leandro, CA).

Statistical Analysis

The statistical significance of differences between the groups and several pathologic factors was assessed using the Kruskal–Wallis test. Overall survival and disease-free survival were calculated with the Kaplan–Meier method and the differences in survival were determined by log-rank analysis. Zero time was the date of pulmonary resection and the terminal event was death attributable to cancer, noncancer event, or unknown causes. Significance was defined as P < 0.05.

RESULTS

Thirty-eight samples of completely resected primary lung adenocarcinoma were analyzed for expression of TSLC1 protein by immunohistochemistry. In normal (nontumor) lung tissues, signals specific for TSLC1 were detected primarily on the lateral plasma membranes of bronchiolar and alveolar epithelial cells as well as at cell-cell boundaries of bronchial and bronchiolar serous gland cells. Although epithelial cells in normal lung tissues expressed TSLC1, its expression level frequently was reduced or lost in samples of primary adenocarcinoma of the lung. Patients were classified into three groups based on the grade of TSLC1 expression in the tumor cells. In Group 1 the percentage of positive cells relative to all tumor cells was > 70% (n = 14 patients) (Fig. 1A), in Group 2 the percentage ranged from 20–70% (n = 10 patients) (Fig. 1B), and in Group 3 the percentage was < 20% (n = 14 patients) (Fig 1C).

Figure 1.

Photomicrographs of representative examples of immunohistochemical staining for TSLC1 in pulmonary adenocarcinoma. Sections were reacted with anti-TSLC1 antibody, colored with aminoethylcarbazole (red), and counterstained with hematoxylin. Patients were classified into 3 groups according to the status of TSLC1 expression in the tumor cells: (A) Group 1: the percentage of positive tumor cells relative to all tumor cells was > 70%; (B) Group 2: the percentage of positive tumor cells was between 20–70%; and (C) Group 3: the percentage of positive tumor cells was < 20% (original magnification ×400).

Next, Western blot analyses were performed to detect specific signals of TSLC1 antigen using these samples and we confirmed the accuracy of the classification made based on immunohistochemical analysis. In other words, we examined whether there were differences between immunohistochemistry and Western blot analysis with regard to the evaluation of TSLC1 expression. The levels of TSLC1 evaluated by immunohistochemistry actually were found to be correlated with those estimated by Western blot analysis both in pulmonary adenocarcinomas and in normal lung tissues (Fig. 2).

Figure 2.

Representative examples of Western blot analyses for TSLC1 in pulmonary adenocarcinoma. Normal (nontumor) lung tissue was loaded in the leftmost lane as a positive control. After blotting with anti-TSLC1 antibody, the membrane was stripped and rehybridized with anti-α-tubulin antibody as a control for protein loading and transfer efficiency. kD: kilodaltons.

The relation between the extent of TSLC1 protein expression in tumor tissues and pathologic parameters including pathologic stage, T classification (tumor size), N classification (lymph node metastasis), lymphatic permeation, and vascular involvement was examined. The patients were evenly distributed into the three groups with regard to the clinical background factors. The expression of TSLC1 was found to be reduced in tumors with a more advanced pathologic stage, higher T classification, higher N classification, more lymphatic permeation, and more vascular involvement, suggesting an inverse correlation between tumor behavior and expression of TSLC1 (Table 1).

Table 1. Pathologic Characteristics of Pulmonary Adenocarcinoma Patients and Levels of TSLC1 Expression
FactorGroup 1 (n = 14)Group 2 (n = 10)Group 3 (n = 14)P value
  1. p Stage: pathologic stage; Ly(−): negative lymphatic permeation; Ly(+): positive lymphatic permeation; V(−): negative vascular involvement; V(+): positive vascular involvement.

p Stage I12310.0004
p Stage II105 
p Stage III178 
T112620.0003
T2129 
T3010 
T4113 
N012310.0005
N1116 
N2167 
Ly(−)12310.0001
Ly(+)2713 
V(−)12310.0001
V(+)2713 

The influence of TSLC1 expression on survival was analyzed. All patients underwent a potentially complete resection of the tumor. Overall follow-up ranged from 12–60 months, with a median of 34 months. The 4-year overall survival rates were 84%, 28%, and 7% in Group 1, Group 2, and Group 3 patients, respectively (Fig. 3). The overall survival rate was found to be significantly better in patients with adenocarcinoma expressing higher TSLC1. In addition, we focused on cancer-specific events in their survival. The 4-year disease-free survival rates were 68%, 28%, and 7% in Group 1, Group 2, and Group 3patients, respectively (Fig. 4). The disease-free survival rate also was significantly better in patients with adenocarcinoma expressing higher TSLC1. These data suggested that decreased expression of the TSLC1 gene in tumor cells was indicative of a poor prognosis. Further investigation of the TSLC1 gene as an independent prognostic indicator are needed.

Figure 3.

Overall survival curves of patients with a curatively resected adenocarcinoma distributed according to the status of TSLC1 expression in the tumor cells.

Figure 4.

Disease-free survival curves of patients with a curatively resected adenocarcinoma distributed according to the status of TSLC1 expression in the tumor cells.

DISCUSSION

Recent advances in tumor cell biology have shed light on the associated mechanics of tumor invasion and metastasis, which are complex processes involving an interaction between the biologic characteristics of the tumor cells and the competence of host defenses. Alterations of cell-cell adhesion molecules may be vital with regard to tumor-cell detachment in tumor invasion and metastasis. E-cadherin, a subclass of the cadherin family, is reported to play a major role in the maintenance of intercellular junctions in epithelial tissues.17, 18 It has been reported that underexpression of E-cadherin or its associated cytoplasmic proteins is closely related to dedifferentiation and invasiveness in many types of epithelial carcinomas.17, 19, 20 The mechanism of E-cadherin reduction in tumor tissues remains unclear. TSLC1 as well as E-cadherin are expressed on the lateral plasma membranes of bronchiolar and alveolar epithelial cells, and may play a physiologic role in cell-cell interactions. The invasiveness of the tumor could be influenced largely by TSLC1 as a cell-cell adhesion molecule because there are some similarities between TSLC1 and E-cadherin with regard to their localization and underexpression in tumor cells.

In the current study, we characterized expression of TSLC1 protein in primary resected lung adenocarcinoma cases. The relative frequency of immunopositive tumor cells was variable and was found to increase as the stage of the tumor advanced, implying a complex control of the production of cell-cell adhesion-regulating proteins and suggesting the possibility that expression of TSLC1 is a late event in the formation of adenocarcinoma. To our knowledge, little is known regarding the clinical and prognostic roles of the TSLC1 gene in patients with lung adenocarcinoma. The most striking finding in the current study was the association between lower TSLC1 protein expression levels and a worse disease-specific survival in patients with curatively resected adenocarcinoma. In addition, TSLC1 expression was found to be correlated inversely with clinicopathologic features indicative of aggressive tumor biology including tumor extent, lymph node involvement, lymphatic permeation, and vascular invasion. These data not only supported the hypothesis that TSLC1 is a tumor suppressor of nonsmall cell lung carcinoma but also demonstrated that TSLC1 gene expression can serve as a predictor of prognosis in patients with lung adenocarcinoma.

To our knowledge to date, pathologic staging remains the most reliable determinant of the prognosis of patients with lung adenocarcinoma and is a main factor influencing the choice of curative treatment. However, our observation added another step to the development of a molecular classification of pulmonary adenocarcinoma and suggested that quantitation of TSLC1 protein expression might help to identify patients at higher risk of recurrence who would benefit from additional therapies to control their disease. If we could evaluate TSLC1 expression using the biopsy materials, the value of TSLC1 as a tumor marker would be greatly enhanced. In these contexts, our finding that TSLC1 expression status is clinically significant in lung adenocarcinoma is of potential interest and requires confirmation.

Acknowledgements

The authors thank Masaharu Kohara (Osaka University Medical School), Wataru Nishio, Toshihiko Sakamoto, Keisuke Hanioka, and Yasuhiro Sakai (Hyogo Medical Center for Adults) for their expert suggestions and assistance during the course of this work.

Ancillary