- Email: [email protected]
Loss of cables protein expression in human non–small cell lung cancer: A tissue microarray study
Loss of Cables Protein Expression in Human Non–Small Cell Lung Cancer: A Tissue Microarray Study DONGFENG TAN, MD, SANDY KIRLEY, MS, QIANG LI, MS, NITHYA RAMNATH, MD, HARRY K. SLOCUM, PHD, JOHN S. BROOKS, MD, CHIN-LEE WU, MD, AND LAWRENCE R. ZUKERBERG, MD Loss of heterozygosity (LOH) on chromosome 18q is common in lung cancer. The genes involved in LOH on 18q in lung cancer have not been well characterized. Cables, a cyclin-dependent kinase (cdk) interacting protein, has recently been identified and mapped to human chromosome 18q11-12. Cables inhibits cell growth and suppresses tumor formation in nude mice, making it a candidate gene for 18q LOH in lung cancer. Little is known regarding Cables protein expression in human non–small cell lung cancer (NSCLC). In this study we examined Cables expression in 163 NSCLC and nonneoplastic lung specimens using tissue microarrays. Strong nuclear staining was present in normal lung and bronchial tissue. We also evaluated the Cables protein expression pattern and its correlation with histopathologic features and with clinical course of NSCLC. The
results of the present study demonstrate for the first time that numerous NSCLCs (45%) lose Cables expression. Furthermore, more adenocarcinomas show a loss of this novel protein than do squamous counterparts. The relationship between tumor histology type and Cables expression appears to be statistically significant (P ⴝ 0.028). Our results suggest that Cables may be involved in the pathogenesis of NSCLC. HUM PATHOL 34:143-149. Copyright 2003, Elsevier Science (USA). All rights reserved. Key words: non–small cell lung cancer, Cables, immunohistochemistry. Abbreviations: cdk, cyclin-dependent kinase; LOH, loss of heterozygosity; NSCLC, non–small cell lung carcinoma.
An estimated 164,000 new cases of lung cancer were diagnosed in the United States in 2002, and lung cancer remains the leading cause of cancer-related deaths in men and women in the United States.1 Non– small cell lung carcinoma (NSCLC) accounts for 80% of all lung cancers. Despite recent advances in surgical, radiation, and medical treatments, the 5-year survival rate of patients with NSCLC remains among the lowest of all major human cancers.1 Our knowledge of biologic abnormalities underlying the development of lung cancer remains quite limited;2,3 thus a better understanding of the molecular and cellular processes that underlie this aggressive malignancy may be of clinical importance. Increasing evidence suggests a connection between aberrations in cell cycle control and cancer. Overexpression of certain cyclins, particularly the Dtype cyclins, contributes to cell transformation.4,5 The p16INK4 gene, an important element in governing the cell cycle, is lost in a considerable number of primary
tumors, and it has been termed the multiple tumorsuppressor 1 gene.6 Similarly, many tumor cell lines proceed through the cell cycle with damaged DNA, suggesting a defect in the regulation of cyclin-dependent kinase (cdk) 2T14/Y15 phosphorylation. The defect could lie in the Wee-1 kinases, cdc25 phosphatases, or other cdk regulatory proteins, such as Cables.7 Cables is a novel cdk-interacting protein that localizes to human chromosome 18q11-12.8-10 We previously found that a Cables-mediated interplay between cdk2 and Wee1 leads to increased cdk2Y15 phosphorylation, inhibitory phosphorylation, and decreased cell growth.10 Loss of Cables expression was observed in about 50% to 60% of human colon and head and neck cancer specimens, and was associated with loss of heterozygosity (LOH) on 18q11.10 There is a high incidence of LOH on chromosome 18q (including the region of Cables) in NSCLC, raising the possibility of a tumor-suppressor gene in this region of chromosome 18.11-14 The aims of the present study were to examine Cables expression and frequency of loss of Cables protein in human NSCLC, to evaluate its association with clinicopathologic parameters, and to assess its possible role in development and prognosis of NSCLC. We studied a cohort of resected primary NSCLCs (n ⫽ 133) using a tissue microarray technique.
From the Department of Pathology, Roswell Park Cancer Institute, State University of New York, Buffalo, NY; Department of Pathology, Massachusetts General Hospital, Boston, MA; and Department of Pathology, University of Pennsylvania, Philadelphia, PA. Accepted for publication November 15, 2002. Supported by a National Cancer Institute Comprehensive Cancer Center support grant (CA16056) to Roswell Park Cancer Institute. Address correspondence and reprint requests to Dongfeng Tan, MD, Department of Pathology, Roswell Park Cancer Institute, State University of New York Buffalo, Elm and Carlton Streets, Buffalo, NY 14263. Copyright 2003, Elsevier Science (USA). All rights reserved. 0046-8177/03/3402-0007$30.00/0 doi:10.1053/hupa.2003.26
MATERIALS AND METHODS Study Population This study is a retrospective analysis of patients diagnosed with NSCLC at Roswell Park Cancer Institute between 1996 and 1999. For inclusion in this study, a patient must have
143
HUMAN PATHOLOGY
Volume 34, No. 2 (February 2003)
had a diagnosis of primary NSCLC, received surgery as the initial treatment modality, had adequate archival tissue for analysis, and had complete clinicopathologic data on record. A total of 133 patients met these criteria. Patients were followed through March 6, 2002. The median patient follow-up time in this sample was 57.2 months. Data collected included age, gender, smoking history, performance status using the Eastern Cooperative Oncology Group scale, date of initial diagnosis, histopathologic diagnosis, grade of tumor differentiation, pathological tumor stage, and date of death from NSCLC or last follow-up. Histological diagnosis and grade of differentiation were determined in accordance with the World Health Organization criteria for lung and pleural tumors.15 Pathological stage was based on the revised international system.16 This study was performed following an Institutional Review Board–approved protocol to investigate molecular markers relevant to lung cancer pathogenesis.
Histological Examination and Construction of NSCLC Tissue Microarrays Hematoxylin and eosin–stained slides of NSCLCs were reviewed for confirmation of histopathologic diagnosis and the adequacy of specimens for immunohistochemical analysis. All carcinomas were reviewed and classified according to the World Health Organization histological classification of lung tumors.15 For most cases (n ⫽ 125), diagnosis was established by examination of conventional hematoxylin and eosin–stained slides. The diagnosis of a minority of cases (spindle cell/sarcomatous squamous cell carcinoma, n ⫽ 5; undifferentiated carcinoma, n ⫽ 2) was confirmed using ancillary techniques, including electron microscopy, immunohistochemistry, and mucin and periodic acid-Schiff stains. Neutral-buffered, formalin-fixed (10% vol/formalin in water, pH 7.4) and paraffin-embedded tissue blocks containing NSCLC and normal lung tissue were retrieved from the Paraffin Archive Resource of the Roswell Park Cancer Institute’s Department of Pathology. To ensure uniformity of sectioning, older blocks were melted and re-embedded using modern plastic cassettes. Areas of tumor and normal tissue elements were identified and marked by an investigator (D.F.T.) for construction of microarrays. Core biopsies of 1 mm diameter were taken from each donor block and arrayed into a recipient block using a microarray system (Beecher Instruments, Silver Spring, MD) as described previously.17 Two separate tissue microarrays were constructed and used in this study. The first array contained 125 cores; the second, 70.
Antibodies and Immunohistochemistry Cables protein antibody purification and IHC were performed as described previously.10 Briefly, glutathione S-transferase fusion proteins were expressed in Escherichia coli and purified with reduced glutathione beads. Polyclonal antisera were raised against a glutathione S-transferase– human Cables protein in rabbits and affinity purified. The antibody recognized a protein of approximately 70,000 kilodaltons in cell lysates on sodium dodecyl sulfate-polyacrylamide gel electrophoresis that comigrated with Cables synthesized in rabbit reticulocyte lyses in vitro and that was recognized by the anti-mouse Cables antisera. Microarray sections were processed within 2 weeks of cutting, to avoid oxidation of antigens. A number of 5-m-thick tissue sections were cut from the microarrays and stained with affinity-purified anti-Cables sera at a 1:200 dilution. Microwave antigen retrieval and avidin-biotin staining methods were used.18 Negative control
sections were immunostained under the same conditions, but substituting preabsorbed antisera and preimmune rabbit antisera for primary antibodies. The immunohistochemistry slides were independently reviewed by 2 investigators (D.F.T. and L.R.Z.). For Cables protein staining, 2 categories were recorded: negative and positive. Only convincing nuclear staining was designated as positive. Conflicting scores (n ⫽ 3) were not included in the final statistical analysis.
Statistical Analysis For all statistical tests, 2 categories in pairs were analyzed as positive versus negative. Categorical variables were analyzed using Fisher’s exact test. Age was analyzed using the Mann-Whitney rank sum test. Follow-up time was calculated using the potential follow-up method. Overall patient survival was calculated from the date of diagnosis to the date of last follow-up (censored) or date of patient death (event). Differences in survival times between patient subgroups were analyzed using the log-rank statistical test. Survival probabilities were calculated using the Kaplan-Meier method. Cox proportional hazards regression analysis was used to determine the association between clinicopathologic variables and overall survival. Statistical significance for model parameters was based on the likelihood ratio test. In all tests, statistical significance was set at 5%.
RESULTS Clinicopathologic Data The clinicopathologic features of the subjects evaluated are summarized in Table 1. The median patient age was 65.1 years (range, 37.5 to 88.1 years). Fifty-eight patients (51%) were male, and 54 (49%) were female. One hundred and one patients (90%) had a smoking history. Seventy-five (67%) patients had a performance score of 0, and 37 (33%) patients had performance scores of 1, 2, or 3. Among the 111 patients with available weight loss information, weight loss was present in 15 (14%). Adenocarincoma (n ⫽ 66) was the primary tumor type (60%) in this patient sample; squamous cell carcinoma (n ⫽ 37) was the second most common type (32%). Fortytwo tumors (37%) had a “well or moderate” histologic grade, and 71 tumors (63%) were graded as “poor or undifferentiated.” Seventy-five patients (66%) had stage I or II disease, and 38 patients (34%) had stage III or IV disease. Cables Expression in Normal Lung Tissue and in NSCLC To evaluate the expression of Cables protein in normal pulmonary tissue, we examined normal lung tissue sections included in the tissue microarrays. Of the 58 nonneoplastic lung tissue cores, 47 could be evaluated. Among these 47 cores, 3 contained hilar lymph nodes, 3 contained bronchial cartilage, 5 contained submucosal glands of bronchus, 5 contained subpleural tissue, 7 contained bronchial mucosa, 11 contained alveolar histiocytes, and 28 contained alveolar epithelium. As expected, all of the nonneoplastic tissue elements expressed Cables protein; although the
144
CABLES PROTEIN EXPRESSION IN NON–SMALL CELL LUNG CANCER (Tan et al)
TABLE 1. Association of Cables Protein Expression With Clinicopathologic Variables Variables Age Sex Smoking history* Performance status* Weight loss* Histology‡
Histological grade§ Stage pN
ⱕmedian ⬎median 0 (Female) 1 (male) 0 (never) 1 (ever) 0 1, 2, 3 0 1 A S AS L BAC, well, mod poor, or ua I, II III, or IV 0 1, 2
Negative (%)
Positive (%)
Total(%)
P
24 (21) 26 (23) 24 (21) 26 (23) 4 (4) 46 (41) 32 (29) 18 (16) 43 (39) 6 (5) 36 (32) 10 (9) 2 (2) 1 (1) 15 (13) 35 (31) 31 (27) 19 (17) 29 (27) 19 (17)
33 (29) 30 (27) 31 (27) 32 (28) 7 (6) 55 (49) 43 (38) 19 (17) 53 (48) 9 (8) 30 (27) 27 (24) 1 (1) 4 (4) 27 (24) 36 (32) 44 (39) 19 (17) 36 (33) 25 (23)
57 (50) 56 (50) 55 (49) 58 (51) 11 (10) 101 (90) 75 (67) 37 (33) 96 (87) 15 (14) 66 (60) 37 (33) 3 (3) 5 (4) 42 (37) 71 (63) 75 (66) 38 (34) 65 (60) 44 (40)
0.64 0.90 0.56 0.55 0.73 0.028
0.16 0.38 0.88
*Two cases lacked data for smoking, weight loss, and performance status. ‡A, adenocarcinoma; S, squamous cell carcinoma; AS, adenosquamous cell carcinoma; L, large-cell carcinoma. Two undifferentiated carcinoma are not included in analysis. §BAC, bronchioalevolar adenocarcinoma. Well, well differentiated; mod, moderately differentiated; poor, poorly differentiated; ua, undifferentiated.
intensity of staining varied among the tissue microarray cores, it generally was constant within the same core. The hilar lymph nodes showed a more diffuse staining pattern. An example of nonneoplastic alveolar epithelium is illustrated in Figure 1A. Of the 137 tumor specimen microcores, 116 could be evaluated. The intensity of staining varied among the tumors, but generally was constant within the same tumor. A few tumors showed focal patchy positive staining, but most tumors were either fully positive or negative. In squamous cell carcinomas, the peripheral areas tended to have stronger staining than the central keratinized areas. Three cases with inconsistent scores were not included in final analysis. Overall, 63 tumors (55%) exhibited positive Cables protein expression (Fig 1B and C), and 50 tumors (45%) exhibited negative expression (Fig 1D). In addition, the expression of Cables protein was also noted in stromal cells (endothelial cells and fibroblasts) and scattered lymphocytes or lymphoid follicles, which served as a positive control in negative cases. Sections incubated with preabsorbed or preimmune rabbit antisera consistently showed no immunostaining. To verify the aforementioned tissue microarray results, whole tissue blocks containing both NSCLC and entrapped nonneoplastic bronchial epithelium were stained with the Cables antibody. The findings from the whole tissue sections were consistent with those from smaller tissue microarray cores. Both positive and negative tumor sections, along with adjacent or entrapped normal tissue, are illustrated in Figure 2.
Association of Cables Protein Expression With Clinicopathologic Parameters The 2 test was used to assess the significance of the association between Cables expression and clinicopathologic variables (Table 1). More adenocarcinomas (54%) showed a loss of Cables expression than did squamous cell carcinomas (27%). The relationship between tumor histology and Cables protein appeared to be statistically significant (P ⫽ 0.028). No other significant relationships were observed between Cables expression and any other clinicopathologic variables. Association of Cables Protein Expression With Overall Survival The estimated survival distributions were calculated using the Kaplan-Meier method. Cox proportional hazards regression analysis was used to derive risk estimates related to overall patient survival for each of the clinicopathologic factors. Overall, those NSCLC patients with loss of Cables staining had a median survival time of 46.4 months, whereas those with positive staining had a median survival of 44.1 months. The difference in survival duration between these 2 groups was not statistically significant (P ⫽ 0.15). In univariate analysis, a highly significant association was seen between patient age, performance status, tumor stage, and survival. In the multivariate model, tumor stage, patient age, and performance status all remained significantly related to survival, whereas no significance was observed between Cables expression and survival. Patients with stage III or IV disease had about 4 times the risk of
145
HUMAN PATHOLOGY
Volume 34, No. 2 (February 2003)
FIGURE 1. Representative tissue microarray core photographs of expression of Cables protein in lung tissue and NSCLC. (A) Nonneoplastic pulmonary alveolar pneumocytes with positive nuclear staining. (B) A squamous cell carcinoma (patient 12) with diffuse nuclear staining. (C) An invasive adenocarcinoma (patient 104) with positive staining of Cables protein. (D) A moderately differentiated adenocarcinoma (patient 48) with negative staining. Note that the stromal cells and infiltrating inflammatory cells are positively stained. (Original magnification ⫻400.)
death from disease than patients with stage I or II (relative risk, 4.02; 95% confidence interval, 2.25 to 7.17; P ⬍0.001). In addition, the patient’s age and performance status and high tumor grade also demonstrated significant value in the prognosis of NSCLC (Table 2). DISCUSSION Various tumor-suppressor genes, including p53, p16, RB, and PTEN, are involved in human lung carcinogenesis.19 In addition, the high incidence of LOH on chromosome 18q in NSCLC points to the presence of 1 or more tumor-suppressor genes on this chromosome arm.11-14 Allelic imbalance of chromosome 18q was detected in up to 60% of NSCLC specimens, including 16 of 28 (57%) with allelic imbalance at 18q11.2.11 In a similar study, 59% of NSCLC-derived cell lines showed
18q11.2 allelic loss, and 1 cell line showed homozygous deletion at this site.12 The same study identified homozygous deletions at 3p14 (FHIT) and 9p21 (p16), where known tumor-suppressor genes are located. Additional evidence for a tumor-suppressor gene at this locus comes from studies showing LOH of 18q11.1q12.3 without distal loss of chromosome 18.20 Several candidate tumor-suppressor genes, including deleted in colon cancer (DCC), Smad4 (DPC4), and Smad2, have been identified in the distal portion of chromosome 18q. DCC was recently shown to be the netrin-1 receptor that directly binds to netrin-1. It is expressed in both normal epithelium and most primary and metastatic cancers. Furthermore DCC-null mice do not develop tumors.21-22 The Smad proteins mediate transforming growth factor- effects and regulate genes involved in cell cycle control. Biallelic inactivation of Smad4 occurs in ⬎60% of pancreas tumors, with few
146
CABLES PROTEIN EXPRESSION IN NON–SMALL CELL LUNG CANCER (Tan et al)
FIGURE 2. Representative whole tissue section photographs of Cables protein expression in NSCLC and entrapped normal bronchial epithelium. (A) The inset shows an entrapped nonneoplastic bronchial epithelium with goblet cells and ciliated cells stained positive for Cables protein, with the surrounding poorly differentiated NSCLC uniformly negative. (B) Both a well-differentiated adenocarcinoma and adjacent bronchial epithelium displayed diffusely positive nuclear staining. (Original magnification ⫻400, ⫻100 [inset].)
mutations identified in the Smad genes in other tumors.23-24 Thus it is likely that chromosome 18q harbors at least 1 proximally located gene involved in human cancer. This is supported by the proximal loss of 18q11-12 without distal 18q loss in some tumors.20 Thus Cables may be the proximal 18q gene involved in human cancer.10 Cables was recently mapped to chromosome 18q11.2-12.1 and was found to inhibit cell growth.10 Furthermore, loss of Cables expression also correlated with LOH on chromosome 18q11-12. Cables appears to be a novel protein that acts as a link or cable between the cdks and nonreceptor tyrosine kinases.8 In proliferating cells, Cables links cdk2 and Wee1 and enhances cdk2Y15 phosphorylation, an inhibitory phosphorylation needed to maintain normal cell cycle checkpoint regulation.10 In yeast, loss of this inhibitory phosphorylation from loss of Wee1 leads to uncontrolled cell growth and mitotic catastrophe.25 In the present study, we were able to study nuclear Cables protein in a large number of lung tissue and NSCLC samples in tissue microarrays using affinitypurified polyclonal antisera. We evaluated the Cables protein expression pattern and its correlation with histopathologic features and with clinical course of NSCLC. The results of this study demonstrate for the first time that (1) a number of NSCLCs lose Cables protein expression, and (2) more adenocarcinomas lose this newly characterized protein than do squamous cell carcinomas. The relationship between tumor histology and Cables protein appears to be statistically significant (P ⫽ 0.028). One concern about tissue microarrays has been the small tissue sample size. In our study, we used the needle core diameter of 1 mm rather than the conventional 0.6 mm, to maximize tissue representation. We also found that Cables was usually diffusely present in
TABLE 2. Relative Risks (RRs) and 95% Confidence Intervals (CIs) From the Analysis of Clinicopathologic Parameters Associated With Survival Relative Risk
95% CI
P
Negative Positive
Ref 1.25
0.74–2.09
0.40
ⱕmedian age ⬎median age 1 (male) 0 (female) 0 (never) 1 (ever) 0 1–3 0 (absent) 1 (present) A S AS L BAC, well, mod poor, ua I, II III or IV 0 1, 2
Ref 1.42 Ref 0.74 Ref 1.04 Ref 2.33 Ref 1.63 Ref 1.31 1.32 1.39 Ref 1.42 Ref 2.72 Ref 2.46
0.99–2.03
0.055
0.52–1.06
0.10
0.56–1.94
0.90
1.63–3.34
⬍0.001*
1.02–2.60
0.039‡
0.88–1.92 0.57–3.04 0.56–3.45
0.18 0.51 0.48
0.98–2.06
0.06
1.90–3.88
⬍0.001*
1.68–3.61
⬍0.001*
Negative Positive
Ref 1.52
0.86–2.68
0.15
ⱕmedian age ⬎median age 0 1–3 BAC, well, mod poor, ua I, II III or IV
Ref 1.93 Ref 2.75 Ref 1.94 Ref 4.02
1.10–3.40
0.023‡
1.53–4.94
0.001*
1.05–3.58
0.035‡
2.25–7.17
⬍0.001*
Variables Univariate analysis Cables Age at diagnosis (years) Sex Smoking history Performance status Weight loss Histology
Histological grade§ Tumor stage pN Multivariate analysis Cables Age at diagnosis (years) Performance status Histological grade Tumor stage
*Statistically significant. ‡P ⬍ 0.05 §BAC, bronchiolveolar adenocarcinoma. Well, well differentiated; mod, moderately differentiated; poor, poorly differentiated; ua, undifferentiated.
147
HUMAN PATHOLOGY
Volume 34, No. 2 (February 2003)
Cables expression was seen more frequently in adenocarcinomas than in squamous cell cancers. These data are consistent with the finding that LOH of chromosome 18q was significantly (P ⫽ 0.036) more frequent in primary lung adenocarcinomas than in squamous cell carcinoma of the lung.13 In general, adenocarcinoma exhibits more clinically aggressive behavior than squamous cell cancer. However, no statistical significance was observed between loss of Cables expression and clinical outcome of patients with NSCLC. This finding correlates with previous findings that clinical features including sex, age, and smoking history were similar in lung cancer cases with or without LOH at chromosome 18q. Furthermore, chromosome 18q LOH was not associated with more or less frequent nodal involvement or with better or worse survival.14
FIGURE 3. Cox regression survival curve of non-small-cell lung cancer patients according to Cables expression: positive versus negative (P ⫽ 0.15).
positive cases. Compared with other tumor markers studied (eg, Her2/neu and E-cadherin),26 Cables protein expression is much easier to assess. It is exclusively confined to the nuclei of tumor cells, rather than in cytoplasm or on cell membrane. Furthermore, tumor marker immunoreactivity in NSCLC has recently proven useful in scanty cytological specimens.27 In addition, as has been demonstrated previously, even focally expressed proteins (eg, chromogranin A) can be assessed using tissue microarrays if a large number of samples are evaluated.28 Therefore, we are confident that the sample cores used in the tissue microarrays in our study are reasonably representative. Cables is a novel protein involved in cell growth control. Loss of Cables may impair the cell cycle and differentiation, which can lead to uncontrolled cell growth and eventually cancer. In the present study, nuclear staining of Cables was detected in normal lung and bronchial tissue. However, 45% (50 of 113) of NSCLCs exhibited loss of nuclear Cables staining. In each of these cases, normal stromal and inflammatory cells were present in and around the negative carcinoma cells to serve as a positive internal control. The percentage loss of Cables expression was slightly lower in our study than in the LOH study (up to 60%).11 Loss of Cables protein expression may not entirely reflect LOH on chromosome 18 at q11-12. The possible relationship between loss of Cables protein and LOH at 18q11-12 in NSCLC is currently under investigation. Nevertheless, lack of expression of Cables in almost 50% of primary lung cancers compared with nonneoplastic counterparts, along with its chromosomal location and role in growth control, suggests that Cables may have a role in the development of these tumors. In this tissue microarray study, analysis of Cables expression or loss of expression with a large number of clinicopathologic variables found a statistically significant association with histologic tumor type. Loss of
Acknowledgment. The authors thank Helen Zou for her expert assistance in the image editing, Joan Natielia and Amy Beck for their assistance in the preparation and processing of tissues, and Weifeng Dong for his expert assistance in the construction of the tissue microarrays.
REFERENCES 1. Greenlee RT, Hill-Harmon BB, Murray T, et al: Cancer statistics, 2001. CA Cancer J Clin 51:16-36, 2002 2. Salgia R, Skarin AT: Molecular abnormalities in lung cancer. J Clin Oncol 16:1207-1217, 1988 3. Pass HI, Mitchell JB, Johnson DH, et al (eds): Lung Cancer: Principles and Practice. Philadelphia, PA, Lippincott Williams & Wilkins, 2000 4. Rosenberg CL, Wong E, Petty EM, et al: PRAD1, a candidate BCL1 oncogene: Mapping and expression in centrocytic lymphoma. Proc Natl Acad Sci U S A, 88:9638-9642, 1991 5. Wang TC, Cardiff RD, Zukerberg L, et al: Mammary hyperplasia and carcinoma in MMTV-cyclin D1 transgenic mice. Nature (Lond) 369:669-671, 1994 6. Bonnetta L: Open questions on p16. Nature (Lond) 370:180, 1994 7. Parsons R: Phosphatase and tumorigenesis. Curr Opin Oncol 10:88-91, 1998 8. Zukerberg LR, Patrick GN, Nikolic M, et al: Cables links cdk5 and c-Abl and facilitates cd5 tyrosine phosphorylation, kinase upregulation, and neurite outgrowth. Neuron 26:633-646, 2000 9. Matsuoka M, Matsuura Y, Semba K, et al: Molecular cloning of a cyclin-like protein associated with cyclin-dependent kinase 3 (cdk3) in vivo. Biochem Biophys Res Commun 273:442-447, 2000 10. Wu CL, Kirley SD, Xiao H, et al: Cables enhances cdk2 tyrosine 15 phosphorylation by Weel, inhibits cell growth, and is lost in many human colon and squamous cancers. Cancer Res 61:73257332, 2001 11. Takei K, Kohno T, Hamada K, et al: A novel tumor suppressor locus on chromosome 18q involved in the development of lung cancer. Cancer Res 58:3700-3705, 1998 12. Girard L, Zochbauer-Muller S, Virmani AK, et al: Genomewide allelotyping of lung cancer identifies new regions of allelic loss, differences between small cell lung cancer and non–small cell lung cancer, and loci clustering. Cancer Res 60:4894-4906, 2000 13. Shiseki M, Kohno T, Nishikawa R, et al: Frequent allelic losses on chromosomes 2q, 18q, and 22q in advanced non–small cell lung carcinoma. Cancer Res 54:5643-5648, 1994 14. Fong KM, Zimmerman PV, Smith PJ: Tumor progression and loss of heterozygosity at 5q and 18q in non–small cell lung cancer. Cancer Res 55:220-223, 1995 15. Travis WD, Colby TV, Corrin B, et al: World Health Organization Classification of Lung and Pleural Tumors. Berlin, Germany, Springer-Verlag, 1999
148
CABLES PROTEIN EXPRESSION IN NON–SMALL CELL LUNG CANCER (Tan et al) 16. Mountain C: Revisions in the international system for staging lung cancer. Chest 111:1710-1717, 1997 17. Kononen J, Bubendorf L, Kallioniemi A, et al: Tissue microarrays for high-throughput molecular profiling of tumor specimens. Nat Med 4:844-847, 1998 18. Ramnath N, Hernandez F, Tan DF, et al: Mdm-2 is an independent predictor of survival in non–small-cell lung cancer. J Clin Oncol 19:259-266, 2001 19. Fong KM, Sekido Y, Minna JD: Molecular pathogenesis of lung cancer. J Thorac Cardiovasc Surg 118:1136-1152, 1999 20. Jones JW, Raval JR, Beals TF, et al: Frequent loss of heterozygosity on chromosome 18q in squamous cell carcinomas. Arch Otolaryngol Head Neck Surg 123:610-614, 1997 21. Kwino-Masu K, Masu M, Hinck L, et al: Deleted in colorectal cancer (DCC) encodes a neutron receptor. Cell 87:175-185, 1996 22. Fazeli A, Dickinson SL, Hermiston ML, et al: A phenotype of mice lacking functional deleted in colorectal cancer (DCC) gene. Nature (Lond) 386:796-804, 1997
23. Hahn SA, Hoque A, Mosaluk CL, et al: Homozygous deletion map at 18q21.1 in pancreatic cancer. Cancer Res 56:490494, 1996 24. Schutte M, Hruban RH, Hedrick L, et al: DPC4 gene in various tumor types. Cancer Res 56:2527-2530, 1996 25. Russell P, Nurse P: Negative regulation of mitosis by wee1⫹, a gene encoding a protein kinase homology. Cell 49:559567, 1987 26. Selvaggi G, Scagliotti GV, Torri V, et al: Her-2/neu overexpression in patients with radically resected non–small cell lung carcinoma. Cancer 94:2669-2674, 2000 27. Harlamert HA, Mira J, Bejarano PA, et al: Thyroid transcription factor-1 and cytokeratins 7 and 20 in pulmonary and breast carcinomas. Acta Cytol 42:1382-1388, 1998 28. Mucci NR, Akdas G, Manely S, et al: Neuroendocrine expression in metastatic prostate cancer: Evaluation of high-throughput tissue microarrays to detect heterogeneous protein expression. HUM PATHOL 31:406-414, 2000
149
results of the present study demonstrate for the first time that numerous NSCLCs (45%) lose Cables expression. Furthermore, more adenocarcinomas show a loss of this novel protein than do squamous counterparts. The relationship between tumor histology type and Cables expression appears to be statistically significant (P ⴝ 0.028). Our results suggest that Cables may be involved in the pathogenesis of NSCLC. HUM PATHOL 34:143-149. Copyright 2003, Elsevier Science (USA). All rights reserved. Key words: non–small cell lung cancer, Cables, immunohistochemistry. Abbreviations: cdk, cyclin-dependent kinase; LOH, loss of heterozygosity; NSCLC, non–small cell lung carcinoma.
An estimated 164,000 new cases of lung cancer were diagnosed in the United States in 2002, and lung cancer remains the leading cause of cancer-related deaths in men and women in the United States.1 Non– small cell lung carcinoma (NSCLC) accounts for 80% of all lung cancers. Despite recent advances in surgical, radiation, and medical treatments, the 5-year survival rate of patients with NSCLC remains among the lowest of all major human cancers.1 Our knowledge of biologic abnormalities underlying the development of lung cancer remains quite limited;2,3 thus a better understanding of the molecular and cellular processes that underlie this aggressive malignancy may be of clinical importance. Increasing evidence suggests a connection between aberrations in cell cycle control and cancer. Overexpression of certain cyclins, particularly the Dtype cyclins, contributes to cell transformation.4,5 The p16INK4 gene, an important element in governing the cell cycle, is lost in a considerable number of primary
tumors, and it has been termed the multiple tumorsuppressor 1 gene.6 Similarly, many tumor cell lines proceed through the cell cycle with damaged DNA, suggesting a defect in the regulation of cyclin-dependent kinase (cdk) 2T14/Y15 phosphorylation. The defect could lie in the Wee-1 kinases, cdc25 phosphatases, or other cdk regulatory proteins, such as Cables.7 Cables is a novel cdk-interacting protein that localizes to human chromosome 18q11-12.8-10 We previously found that a Cables-mediated interplay between cdk2 and Wee1 leads to increased cdk2Y15 phosphorylation, inhibitory phosphorylation, and decreased cell growth.10 Loss of Cables expression was observed in about 50% to 60% of human colon and head and neck cancer specimens, and was associated with loss of heterozygosity (LOH) on 18q11.10 There is a high incidence of LOH on chromosome 18q (including the region of Cables) in NSCLC, raising the possibility of a tumor-suppressor gene in this region of chromosome 18.11-14 The aims of the present study were to examine Cables expression and frequency of loss of Cables protein in human NSCLC, to evaluate its association with clinicopathologic parameters, and to assess its possible role in development and prognosis of NSCLC. We studied a cohort of resected primary NSCLCs (n ⫽ 133) using a tissue microarray technique.
From the Department of Pathology, Roswell Park Cancer Institute, State University of New York, Buffalo, NY; Department of Pathology, Massachusetts General Hospital, Boston, MA; and Department of Pathology, University of Pennsylvania, Philadelphia, PA. Accepted for publication November 15, 2002. Supported by a National Cancer Institute Comprehensive Cancer Center support grant (CA16056) to Roswell Park Cancer Institute. Address correspondence and reprint requests to Dongfeng Tan, MD, Department of Pathology, Roswell Park Cancer Institute, State University of New York Buffalo, Elm and Carlton Streets, Buffalo, NY 14263. Copyright 2003, Elsevier Science (USA). All rights reserved. 0046-8177/03/3402-0007$30.00/0 doi:10.1053/hupa.2003.26
MATERIALS AND METHODS Study Population This study is a retrospective analysis of patients diagnosed with NSCLC at Roswell Park Cancer Institute between 1996 and 1999. For inclusion in this study, a patient must have
143
HUMAN PATHOLOGY
Volume 34, No. 2 (February 2003)
had a diagnosis of primary NSCLC, received surgery as the initial treatment modality, had adequate archival tissue for analysis, and had complete clinicopathologic data on record. A total of 133 patients met these criteria. Patients were followed through March 6, 2002. The median patient follow-up time in this sample was 57.2 months. Data collected included age, gender, smoking history, performance status using the Eastern Cooperative Oncology Group scale, date of initial diagnosis, histopathologic diagnosis, grade of tumor differentiation, pathological tumor stage, and date of death from NSCLC or last follow-up. Histological diagnosis and grade of differentiation were determined in accordance with the World Health Organization criteria for lung and pleural tumors.15 Pathological stage was based on the revised international system.16 This study was performed following an Institutional Review Board–approved protocol to investigate molecular markers relevant to lung cancer pathogenesis.
Histological Examination and Construction of NSCLC Tissue Microarrays Hematoxylin and eosin–stained slides of NSCLCs were reviewed for confirmation of histopathologic diagnosis and the adequacy of specimens for immunohistochemical analysis. All carcinomas were reviewed and classified according to the World Health Organization histological classification of lung tumors.15 For most cases (n ⫽ 125), diagnosis was established by examination of conventional hematoxylin and eosin–stained slides. The diagnosis of a minority of cases (spindle cell/sarcomatous squamous cell carcinoma, n ⫽ 5; undifferentiated carcinoma, n ⫽ 2) was confirmed using ancillary techniques, including electron microscopy, immunohistochemistry, and mucin and periodic acid-Schiff stains. Neutral-buffered, formalin-fixed (10% vol/formalin in water, pH 7.4) and paraffin-embedded tissue blocks containing NSCLC and normal lung tissue were retrieved from the Paraffin Archive Resource of the Roswell Park Cancer Institute’s Department of Pathology. To ensure uniformity of sectioning, older blocks were melted and re-embedded using modern plastic cassettes. Areas of tumor and normal tissue elements were identified and marked by an investigator (D.F.T.) for construction of microarrays. Core biopsies of 1 mm diameter were taken from each donor block and arrayed into a recipient block using a microarray system (Beecher Instruments, Silver Spring, MD) as described previously.17 Two separate tissue microarrays were constructed and used in this study. The first array contained 125 cores; the second, 70.
Antibodies and Immunohistochemistry Cables protein antibody purification and IHC were performed as described previously.10 Briefly, glutathione S-transferase fusion proteins were expressed in Escherichia coli and purified with reduced glutathione beads. Polyclonal antisera were raised against a glutathione S-transferase– human Cables protein in rabbits and affinity purified. The antibody recognized a protein of approximately 70,000 kilodaltons in cell lysates on sodium dodecyl sulfate-polyacrylamide gel electrophoresis that comigrated with Cables synthesized in rabbit reticulocyte lyses in vitro and that was recognized by the anti-mouse Cables antisera. Microarray sections were processed within 2 weeks of cutting, to avoid oxidation of antigens. A number of 5-m-thick tissue sections were cut from the microarrays and stained with affinity-purified anti-Cables sera at a 1:200 dilution. Microwave antigen retrieval and avidin-biotin staining methods were used.18 Negative control
sections were immunostained under the same conditions, but substituting preabsorbed antisera and preimmune rabbit antisera for primary antibodies. The immunohistochemistry slides were independently reviewed by 2 investigators (D.F.T. and L.R.Z.). For Cables protein staining, 2 categories were recorded: negative and positive. Only convincing nuclear staining was designated as positive. Conflicting scores (n ⫽ 3) were not included in the final statistical analysis.
Statistical Analysis For all statistical tests, 2 categories in pairs were analyzed as positive versus negative. Categorical variables were analyzed using Fisher’s exact test. Age was analyzed using the Mann-Whitney rank sum test. Follow-up time was calculated using the potential follow-up method. Overall patient survival was calculated from the date of diagnosis to the date of last follow-up (censored) or date of patient death (event). Differences in survival times between patient subgroups were analyzed using the log-rank statistical test. Survival probabilities were calculated using the Kaplan-Meier method. Cox proportional hazards regression analysis was used to determine the association between clinicopathologic variables and overall survival. Statistical significance for model parameters was based on the likelihood ratio test. In all tests, statistical significance was set at 5%.
RESULTS Clinicopathologic Data The clinicopathologic features of the subjects evaluated are summarized in Table 1. The median patient age was 65.1 years (range, 37.5 to 88.1 years). Fifty-eight patients (51%) were male, and 54 (49%) were female. One hundred and one patients (90%) had a smoking history. Seventy-five (67%) patients had a performance score of 0, and 37 (33%) patients had performance scores of 1, 2, or 3. Among the 111 patients with available weight loss information, weight loss was present in 15 (14%). Adenocarincoma (n ⫽ 66) was the primary tumor type (60%) in this patient sample; squamous cell carcinoma (n ⫽ 37) was the second most common type (32%). Fortytwo tumors (37%) had a “well or moderate” histologic grade, and 71 tumors (63%) were graded as “poor or undifferentiated.” Seventy-five patients (66%) had stage I or II disease, and 38 patients (34%) had stage III or IV disease. Cables Expression in Normal Lung Tissue and in NSCLC To evaluate the expression of Cables protein in normal pulmonary tissue, we examined normal lung tissue sections included in the tissue microarrays. Of the 58 nonneoplastic lung tissue cores, 47 could be evaluated. Among these 47 cores, 3 contained hilar lymph nodes, 3 contained bronchial cartilage, 5 contained submucosal glands of bronchus, 5 contained subpleural tissue, 7 contained bronchial mucosa, 11 contained alveolar histiocytes, and 28 contained alveolar epithelium. As expected, all of the nonneoplastic tissue elements expressed Cables protein; although the
144
CABLES PROTEIN EXPRESSION IN NON–SMALL CELL LUNG CANCER (Tan et al)
TABLE 1. Association of Cables Protein Expression With Clinicopathologic Variables Variables Age Sex Smoking history* Performance status* Weight loss* Histology‡
Histological grade§ Stage pN
ⱕmedian ⬎median 0 (Female) 1 (male) 0 (never) 1 (ever) 0 1, 2, 3 0 1 A S AS L BAC, well, mod poor, or ua I, II III, or IV 0 1, 2
Negative (%)
Positive (%)
Total(%)
P
24 (21) 26 (23) 24 (21) 26 (23) 4 (4) 46 (41) 32 (29) 18 (16) 43 (39) 6 (5) 36 (32) 10 (9) 2 (2) 1 (1) 15 (13) 35 (31) 31 (27) 19 (17) 29 (27) 19 (17)
33 (29) 30 (27) 31 (27) 32 (28) 7 (6) 55 (49) 43 (38) 19 (17) 53 (48) 9 (8) 30 (27) 27 (24) 1 (1) 4 (4) 27 (24) 36 (32) 44 (39) 19 (17) 36 (33) 25 (23)
57 (50) 56 (50) 55 (49) 58 (51) 11 (10) 101 (90) 75 (67) 37 (33) 96 (87) 15 (14) 66 (60) 37 (33) 3 (3) 5 (4) 42 (37) 71 (63) 75 (66) 38 (34) 65 (60) 44 (40)
0.64 0.90 0.56 0.55 0.73 0.028
0.16 0.38 0.88
*Two cases lacked data for smoking, weight loss, and performance status. ‡A, adenocarcinoma; S, squamous cell carcinoma; AS, adenosquamous cell carcinoma; L, large-cell carcinoma. Two undifferentiated carcinoma are not included in analysis. §BAC, bronchioalevolar adenocarcinoma. Well, well differentiated; mod, moderately differentiated; poor, poorly differentiated; ua, undifferentiated.
intensity of staining varied among the tissue microarray cores, it generally was constant within the same core. The hilar lymph nodes showed a more diffuse staining pattern. An example of nonneoplastic alveolar epithelium is illustrated in Figure 1A. Of the 137 tumor specimen microcores, 116 could be evaluated. The intensity of staining varied among the tumors, but generally was constant within the same tumor. A few tumors showed focal patchy positive staining, but most tumors were either fully positive or negative. In squamous cell carcinomas, the peripheral areas tended to have stronger staining than the central keratinized areas. Three cases with inconsistent scores were not included in final analysis. Overall, 63 tumors (55%) exhibited positive Cables protein expression (Fig 1B and C), and 50 tumors (45%) exhibited negative expression (Fig 1D). In addition, the expression of Cables protein was also noted in stromal cells (endothelial cells and fibroblasts) and scattered lymphocytes or lymphoid follicles, which served as a positive control in negative cases. Sections incubated with preabsorbed or preimmune rabbit antisera consistently showed no immunostaining. To verify the aforementioned tissue microarray results, whole tissue blocks containing both NSCLC and entrapped nonneoplastic bronchial epithelium were stained with the Cables antibody. The findings from the whole tissue sections were consistent with those from smaller tissue microarray cores. Both positive and negative tumor sections, along with adjacent or entrapped normal tissue, are illustrated in Figure 2.
Association of Cables Protein Expression With Clinicopathologic Parameters The 2 test was used to assess the significance of the association between Cables expression and clinicopathologic variables (Table 1). More adenocarcinomas (54%) showed a loss of Cables expression than did squamous cell carcinomas (27%). The relationship between tumor histology and Cables protein appeared to be statistically significant (P ⫽ 0.028). No other significant relationships were observed between Cables expression and any other clinicopathologic variables. Association of Cables Protein Expression With Overall Survival The estimated survival distributions were calculated using the Kaplan-Meier method. Cox proportional hazards regression analysis was used to derive risk estimates related to overall patient survival for each of the clinicopathologic factors. Overall, those NSCLC patients with loss of Cables staining had a median survival time of 46.4 months, whereas those with positive staining had a median survival of 44.1 months. The difference in survival duration between these 2 groups was not statistically significant (P ⫽ 0.15). In univariate analysis, a highly significant association was seen between patient age, performance status, tumor stage, and survival. In the multivariate model, tumor stage, patient age, and performance status all remained significantly related to survival, whereas no significance was observed between Cables expression and survival. Patients with stage III or IV disease had about 4 times the risk of
145
HUMAN PATHOLOGY
Volume 34, No. 2 (February 2003)
FIGURE 1. Representative tissue microarray core photographs of expression of Cables protein in lung tissue and NSCLC. (A) Nonneoplastic pulmonary alveolar pneumocytes with positive nuclear staining. (B) A squamous cell carcinoma (patient 12) with diffuse nuclear staining. (C) An invasive adenocarcinoma (patient 104) with positive staining of Cables protein. (D) A moderately differentiated adenocarcinoma (patient 48) with negative staining. Note that the stromal cells and infiltrating inflammatory cells are positively stained. (Original magnification ⫻400.)
death from disease than patients with stage I or II (relative risk, 4.02; 95% confidence interval, 2.25 to 7.17; P ⬍0.001). In addition, the patient’s age and performance status and high tumor grade also demonstrated significant value in the prognosis of NSCLC (Table 2). DISCUSSION Various tumor-suppressor genes, including p53, p16, RB, and PTEN, are involved in human lung carcinogenesis.19 In addition, the high incidence of LOH on chromosome 18q in NSCLC points to the presence of 1 or more tumor-suppressor genes on this chromosome arm.11-14 Allelic imbalance of chromosome 18q was detected in up to 60% of NSCLC specimens, including 16 of 28 (57%) with allelic imbalance at 18q11.2.11 In a similar study, 59% of NSCLC-derived cell lines showed
18q11.2 allelic loss, and 1 cell line showed homozygous deletion at this site.12 The same study identified homozygous deletions at 3p14 (FHIT) and 9p21 (p16), where known tumor-suppressor genes are located. Additional evidence for a tumor-suppressor gene at this locus comes from studies showing LOH of 18q11.1q12.3 without distal loss of chromosome 18.20 Several candidate tumor-suppressor genes, including deleted in colon cancer (DCC), Smad4 (DPC4), and Smad2, have been identified in the distal portion of chromosome 18q. DCC was recently shown to be the netrin-1 receptor that directly binds to netrin-1. It is expressed in both normal epithelium and most primary and metastatic cancers. Furthermore DCC-null mice do not develop tumors.21-22 The Smad proteins mediate transforming growth factor- effects and regulate genes involved in cell cycle control. Biallelic inactivation of Smad4 occurs in ⬎60% of pancreas tumors, with few
146
CABLES PROTEIN EXPRESSION IN NON–SMALL CELL LUNG CANCER (Tan et al)
FIGURE 2. Representative whole tissue section photographs of Cables protein expression in NSCLC and entrapped normal bronchial epithelium. (A) The inset shows an entrapped nonneoplastic bronchial epithelium with goblet cells and ciliated cells stained positive for Cables protein, with the surrounding poorly differentiated NSCLC uniformly negative. (B) Both a well-differentiated adenocarcinoma and adjacent bronchial epithelium displayed diffusely positive nuclear staining. (Original magnification ⫻400, ⫻100 [inset].)
mutations identified in the Smad genes in other tumors.23-24 Thus it is likely that chromosome 18q harbors at least 1 proximally located gene involved in human cancer. This is supported by the proximal loss of 18q11-12 without distal 18q loss in some tumors.20 Thus Cables may be the proximal 18q gene involved in human cancer.10 Cables was recently mapped to chromosome 18q11.2-12.1 and was found to inhibit cell growth.10 Furthermore, loss of Cables expression also correlated with LOH on chromosome 18q11-12. Cables appears to be a novel protein that acts as a link or cable between the cdks and nonreceptor tyrosine kinases.8 In proliferating cells, Cables links cdk2 and Wee1 and enhances cdk2Y15 phosphorylation, an inhibitory phosphorylation needed to maintain normal cell cycle checkpoint regulation.10 In yeast, loss of this inhibitory phosphorylation from loss of Wee1 leads to uncontrolled cell growth and mitotic catastrophe.25 In the present study, we were able to study nuclear Cables protein in a large number of lung tissue and NSCLC samples in tissue microarrays using affinitypurified polyclonal antisera. We evaluated the Cables protein expression pattern and its correlation with histopathologic features and with clinical course of NSCLC. The results of this study demonstrate for the first time that (1) a number of NSCLCs lose Cables protein expression, and (2) more adenocarcinomas lose this newly characterized protein than do squamous cell carcinomas. The relationship between tumor histology and Cables protein appears to be statistically significant (P ⫽ 0.028). One concern about tissue microarrays has been the small tissue sample size. In our study, we used the needle core diameter of 1 mm rather than the conventional 0.6 mm, to maximize tissue representation. We also found that Cables was usually diffusely present in
TABLE 2. Relative Risks (RRs) and 95% Confidence Intervals (CIs) From the Analysis of Clinicopathologic Parameters Associated With Survival Relative Risk
95% CI
P
Negative Positive
Ref 1.25
0.74–2.09
0.40
ⱕmedian age ⬎median age 1 (male) 0 (female) 0 (never) 1 (ever) 0 1–3 0 (absent) 1 (present) A S AS L BAC, well, mod poor, ua I, II III or IV 0 1, 2
Ref 1.42 Ref 0.74 Ref 1.04 Ref 2.33 Ref 1.63 Ref 1.31 1.32 1.39 Ref 1.42 Ref 2.72 Ref 2.46
0.99–2.03
0.055
0.52–1.06
0.10
0.56–1.94
0.90
1.63–3.34
⬍0.001*
1.02–2.60
0.039‡
0.88–1.92 0.57–3.04 0.56–3.45
0.18 0.51 0.48
0.98–2.06
0.06
1.90–3.88
⬍0.001*
1.68–3.61
⬍0.001*
Negative Positive
Ref 1.52
0.86–2.68
0.15
ⱕmedian age ⬎median age 0 1–3 BAC, well, mod poor, ua I, II III or IV
Ref 1.93 Ref 2.75 Ref 1.94 Ref 4.02
1.10–3.40
0.023‡
1.53–4.94
0.001*
1.05–3.58
0.035‡
2.25–7.17
⬍0.001*
Variables Univariate analysis Cables Age at diagnosis (years) Sex Smoking history Performance status Weight loss Histology
Histological grade§ Tumor stage pN Multivariate analysis Cables Age at diagnosis (years) Performance status Histological grade Tumor stage
*Statistically significant. ‡P ⬍ 0.05 §BAC, bronchiolveolar adenocarcinoma. Well, well differentiated; mod, moderately differentiated; poor, poorly differentiated; ua, undifferentiated.
147
HUMAN PATHOLOGY
Volume 34, No. 2 (February 2003)
Cables expression was seen more frequently in adenocarcinomas than in squamous cell cancers. These data are consistent with the finding that LOH of chromosome 18q was significantly (P ⫽ 0.036) more frequent in primary lung adenocarcinomas than in squamous cell carcinoma of the lung.13 In general, adenocarcinoma exhibits more clinically aggressive behavior than squamous cell cancer. However, no statistical significance was observed between loss of Cables expression and clinical outcome of patients with NSCLC. This finding correlates with previous findings that clinical features including sex, age, and smoking history were similar in lung cancer cases with or without LOH at chromosome 18q. Furthermore, chromosome 18q LOH was not associated with more or less frequent nodal involvement or with better or worse survival.14
FIGURE 3. Cox regression survival curve of non-small-cell lung cancer patients according to Cables expression: positive versus negative (P ⫽ 0.15).
positive cases. Compared with other tumor markers studied (eg, Her2/neu and E-cadherin),26 Cables protein expression is much easier to assess. It is exclusively confined to the nuclei of tumor cells, rather than in cytoplasm or on cell membrane. Furthermore, tumor marker immunoreactivity in NSCLC has recently proven useful in scanty cytological specimens.27 In addition, as has been demonstrated previously, even focally expressed proteins (eg, chromogranin A) can be assessed using tissue microarrays if a large number of samples are evaluated.28 Therefore, we are confident that the sample cores used in the tissue microarrays in our study are reasonably representative. Cables is a novel protein involved in cell growth control. Loss of Cables may impair the cell cycle and differentiation, which can lead to uncontrolled cell growth and eventually cancer. In the present study, nuclear staining of Cables was detected in normal lung and bronchial tissue. However, 45% (50 of 113) of NSCLCs exhibited loss of nuclear Cables staining. In each of these cases, normal stromal and inflammatory cells were present in and around the negative carcinoma cells to serve as a positive internal control. The percentage loss of Cables expression was slightly lower in our study than in the LOH study (up to 60%).11 Loss of Cables protein expression may not entirely reflect LOH on chromosome 18 at q11-12. The possible relationship between loss of Cables protein and LOH at 18q11-12 in NSCLC is currently under investigation. Nevertheless, lack of expression of Cables in almost 50% of primary lung cancers compared with nonneoplastic counterparts, along with its chromosomal location and role in growth control, suggests that Cables may have a role in the development of these tumors. In this tissue microarray study, analysis of Cables expression or loss of expression with a large number of clinicopathologic variables found a statistically significant association with histologic tumor type. Loss of
Acknowledgment. The authors thank Helen Zou for her expert assistance in the image editing, Joan Natielia and Amy Beck for their assistance in the preparation and processing of tissues, and Weifeng Dong for his expert assistance in the construction of the tissue microarrays.
REFERENCES 1. Greenlee RT, Hill-Harmon BB, Murray T, et al: Cancer statistics, 2001. CA Cancer J Clin 51:16-36, 2002 2. Salgia R, Skarin AT: Molecular abnormalities in lung cancer. J Clin Oncol 16:1207-1217, 1988 3. Pass HI, Mitchell JB, Johnson DH, et al (eds): Lung Cancer: Principles and Practice. Philadelphia, PA, Lippincott Williams & Wilkins, 2000 4. Rosenberg CL, Wong E, Petty EM, et al: PRAD1, a candidate BCL1 oncogene: Mapping and expression in centrocytic lymphoma. Proc Natl Acad Sci U S A, 88:9638-9642, 1991 5. Wang TC, Cardiff RD, Zukerberg L, et al: Mammary hyperplasia and carcinoma in MMTV-cyclin D1 transgenic mice. Nature (Lond) 369:669-671, 1994 6. Bonnetta L: Open questions on p16. Nature (Lond) 370:180, 1994 7. Parsons R: Phosphatase and tumorigenesis. Curr Opin Oncol 10:88-91, 1998 8. Zukerberg LR, Patrick GN, Nikolic M, et al: Cables links cdk5 and c-Abl and facilitates cd5 tyrosine phosphorylation, kinase upregulation, and neurite outgrowth. Neuron 26:633-646, 2000 9. Matsuoka M, Matsuura Y, Semba K, et al: Molecular cloning of a cyclin-like protein associated with cyclin-dependent kinase 3 (cdk3) in vivo. Biochem Biophys Res Commun 273:442-447, 2000 10. Wu CL, Kirley SD, Xiao H, et al: Cables enhances cdk2 tyrosine 15 phosphorylation by Weel, inhibits cell growth, and is lost in many human colon and squamous cancers. Cancer Res 61:73257332, 2001 11. Takei K, Kohno T, Hamada K, et al: A novel tumor suppressor locus on chromosome 18q involved in the development of lung cancer. Cancer Res 58:3700-3705, 1998 12. Girard L, Zochbauer-Muller S, Virmani AK, et al: Genomewide allelotyping of lung cancer identifies new regions of allelic loss, differences between small cell lung cancer and non–small cell lung cancer, and loci clustering. Cancer Res 60:4894-4906, 2000 13. Shiseki M, Kohno T, Nishikawa R, et al: Frequent allelic losses on chromosomes 2q, 18q, and 22q in advanced non–small cell lung carcinoma. Cancer Res 54:5643-5648, 1994 14. Fong KM, Zimmerman PV, Smith PJ: Tumor progression and loss of heterozygosity at 5q and 18q in non–small cell lung cancer. Cancer Res 55:220-223, 1995 15. Travis WD, Colby TV, Corrin B, et al: World Health Organization Classification of Lung and Pleural Tumors. Berlin, Germany, Springer-Verlag, 1999
148
CABLES PROTEIN EXPRESSION IN NON–SMALL CELL LUNG CANCER (Tan et al) 16. Mountain C: Revisions in the international system for staging lung cancer. Chest 111:1710-1717, 1997 17. Kononen J, Bubendorf L, Kallioniemi A, et al: Tissue microarrays for high-throughput molecular profiling of tumor specimens. Nat Med 4:844-847, 1998 18. Ramnath N, Hernandez F, Tan DF, et al: Mdm-2 is an independent predictor of survival in non–small-cell lung cancer. J Clin Oncol 19:259-266, 2001 19. Fong KM, Sekido Y, Minna JD: Molecular pathogenesis of lung cancer. J Thorac Cardiovasc Surg 118:1136-1152, 1999 20. Jones JW, Raval JR, Beals TF, et al: Frequent loss of heterozygosity on chromosome 18q in squamous cell carcinomas. Arch Otolaryngol Head Neck Surg 123:610-614, 1997 21. Kwino-Masu K, Masu M, Hinck L, et al: Deleted in colorectal cancer (DCC) encodes a neutron receptor. Cell 87:175-185, 1996 22. Fazeli A, Dickinson SL, Hermiston ML, et al: A phenotype of mice lacking functional deleted in colorectal cancer (DCC) gene. Nature (Lond) 386:796-804, 1997
23. Hahn SA, Hoque A, Mosaluk CL, et al: Homozygous deletion map at 18q21.1 in pancreatic cancer. Cancer Res 56:490494, 1996 24. Schutte M, Hruban RH, Hedrick L, et al: DPC4 gene in various tumor types. Cancer Res 56:2527-2530, 1996 25. Russell P, Nurse P: Negative regulation of mitosis by wee1⫹, a gene encoding a protein kinase homology. Cell 49:559567, 1987 26. Selvaggi G, Scagliotti GV, Torri V, et al: Her-2/neu overexpression in patients with radically resected non–small cell lung carcinoma. Cancer 94:2669-2674, 2000 27. Harlamert HA, Mira J, Bejarano PA, et al: Thyroid transcription factor-1 and cytokeratins 7 and 20 in pulmonary and breast carcinomas. Acta Cytol 42:1382-1388, 1998 28. Mucci NR, Akdas G, Manely S, et al: Neuroendocrine expression in metastatic prostate cancer: Evaluation of high-throughput tissue microarrays to detect heterogeneous protein expression. HUM PATHOL 31:406-414, 2000
149