Published ahead of print on August 18, 2004, doi:10.1164/rccm.200404-487OC
© 2004 American Thoracic Society doi: 10.1164/rccm.200404-487OC
Early Involvement of the Phosphatidylinositol 3-Kinase/Akt Pathway in Lung Cancer ProgressionDivision of Allergy, Pulmonary and Critical Care Medicine; Department of Medicine, Biostatistics Shared Resource; Tissue Informatics Shared Resource; Department of Pathology; and Department of Surgery, The Vanderbilt Ingram Comprehensive Cancer Center, Vanderbilt University School of Medicine; and VA Medical Center, Nashville, Tennessee Correspondence and requests for reprints should be addressed to Pierre P. Massion, M.D., Division of Allergy, Pulmonary and Critical Care Medicine, Vanderbilt University Medical Center, the Vanderbilt-Ingram Comprehensive Cancer Center, 2220 Pierce Avenue, PRB 640, Nashville, TN 37232-6838. E-mail: pierre.massion{at}vanderbilt.edu
Signaling through the phosphatidylinositol 3-kinase (PI3-kinase) pathway has been associated with lung tumorigenesis. We examined the association between gene copy number of the PI3-kinase catalytic subunit  (PIK3CA) and phosphorylated Akt expression in invasive and preinvasive lung cancers. We sought to determine at what stage of tumor development gene copy number increase or phosphorylated Akt overexpression might affect tumor development. We assessed PIK3CA gene copy number by fluorescence in situ hybridization and expression of phosphorylated Akt by immunohistochemistry in 242 invasive and 43 preinvasive lung cancers and correlated our findings with clinical outcome. The PIK3CA was amplified in 70% of squamous carcinomas, 38% of large cell carcinomas, 19% of adenocarcinomas, and 67% of small cell lung cancers. Phosphorylated Akt overexpression was frequently observed, and strongly so in 12 to 17% of lung cancers depending on nuclear or cytoplasmic localization. Neither PIK3CA gene copy number nor phosphorylated Akt protein expression had prognostic significance. In preinvasive lesions, amplification of the PIK3CA and overexpression of phosphorylated Akt were associated with severe dysplasia and each other. These observations suggest frequent and early involvement of the PI3-kinase pathway in lung cancer.
Key Words: amplification • PKB • preinvasive • tissue microarray • tumorigenesis
Genomic abnormalities represent one major signature of neoplastic transformation and tumor progression (1). Patterns of chromosome copy number (CN) abnormalities in lung cancer have recently been identified using comparative genomic hybridization (CGH) analysis (2–4). Using array CGH, we recently identified genetic amplification on chromosome 3q26–3qter as the most common genomic abnormality in squamous carcinoma of the lung (5). The gene coding for the phosphatidylinositol 3-kinase (PI3-kinase) catalytic subunit PIK3CA is activated by a series of cell surface tyrosine kinase receptors such as insulin growth factor receptor and platelet-derived growth factor receptor (6). Upon growth factor receptor activation, PIK3CA binds to its heterodimer, p85, and promotes the phosphorylation of Akt at Serine 473 and/or Threonine 308. Akt then activates a series of cancer-related functions such as cell proliferation, cell migration, invasion, and decrease of cell death in human tumors (for review see Reference 7). Although the PIK3CA gene has been identified as an oncogene in ovarian (8), cervical (9), and head and neck cancer (10), its exact role in lung cancer development remains unclear. We have recently shown that PIK3CA gene copy number correlates with increased Akt activity in squamous carcinoma of the lung (5), implicating the PI3-kinase pathway in lung tumorigenesis. Akt is also known to be activated in airway epithelial cells in vitro after exposure to growth factors such as epidermal growth factor as well as tobacco-related chemicals such as nicotine and NNK (11). In addition, phosphorylated Akt (pAkt) has recently been shown to be overexpressed in 67% of non–small cell lung cancers (NSCLCs) (12) and in preinvasive lesions (13). Finally, recent attempts to interfere with the PI3-kinase/Akt pathway have shown some promise in lung cancer prevention (14). All of these lines of evidence strongly indicate a critical role for the PI3-kinase/Akt pathway in lung cancer progression. Because the PI3-kinase/Akt pathway may have a critical role in lung cancer development, and given that molecular studies of invasive and preinvasive carcinomas of the lung may provide information about progression, staging or recurrence of disease, we asked the following three questions: (1) What is the prevalence of increased PIK3CA gene CN in lung cancer? (2) What is the association between PIK3CA gene CN and phosphorylation of Akt in lung cancer? (3) At what point does PIK3CA gene amplification or pAkt overexpression take place during lung cancer development? Some of the results of these studies have been previously reported in abstract form (15, 16).
Tissue Samples and Tissue Microarrays Paraffin-embedded formalin-fixed tissues (including preinvasive lesions) were obtained from the archives of the pathology department at Vanderbilt University and the Department of Veterans Affairs Medical Center (VAMC, Nashville, TN). Fresh tissues, including lung tumors and normal lung, were obtained from surgical specimens through the Specialized Program of Research Excellence (SPORE) in lung cancer at Vanderbilt. Additional preinvasive lesions were selected from the archives of the pathology department at the University of California San Francisco. The study was approved by the local Institutional Review Boards for all institutions involved. Tissue microarrays of NSCLCs were prepared from paraffin blocks following the methods described by Kononen and colleagues (17) and reported earlier (18). There are 242 tumors represented on the arrays. Hematoxylin and eosin (H&E) stained sections from all tissue blocks were reviewed by our pathologists (A.L.G., J.J.A.).
Fluorescence In Situ Hybridization
Immunohistochemistry
Statistical Analysis
We studied 242 lung cancers on a total of four tissue microarrays. Each lung tumor was assayed in triplicate. The clinical data is summarized in Table 1. We also studied 43 preinvasive lesions of the lung, 20 low-grade and 23 high-grade.
Amplification of PIK3CA Gene CN in Lung Cancer We analyzed PIK3CA gene CN in lung cancer by performing dual FISH analysis on four tissue microarrays. We obtained FISH data on 55% of the cases based on the success of the hybridization and the imperfect representation of all tumors in the tissue microarrays. Unsuccessful FISH hybridizations were likely related to extensive formalin-fixation of archival materials. Overall, PIK3CA was amplified (CN ratio PIK3CA/Cent3 > 1.5) in 43% (57/132) of lung tumors analyzed by dual color FISH (Table 2). When looking at PIK3CA CN ratios by histologic groups, we found amplification in 70% (32/46) of squamous cell carcinomas, 19% (9/47) of adenocarcinomas, 38% (6/16) of large cell carcinomas, and 67% (8/12) of small cell lung cancers (SCLCs), 0/4 carcinoid, 0/2 adenosquamous, 0/1 adenocystic, 0/1 sarcoma, and 2/3 NSCLCs showed amplification. A higher degree of amplification (CN ratio > 2) was found in 33% of the tumors, but was most predominant in squamous carcinomas (Table 2) where high level amplification accounted for 61% (28/46) of all amplifications. The CN status for PIK3CA was not associated with age (p = 0.1043), smoking duration (p = 0.1146), pack-years (p = 0.4278), nodal status (p = 0.3437), stage (p = 0.3066), or survival (p = 0.6281).
Expression of pAkt in Lung Cancer and Association with PIK3CA Amplification Because of the critical role of an activated Akt pathway in cancer and the possible association with upstream amplification of PIK3CA, we analyzed the level of expression of pAkt in NSCLC by immunohistochemistry. Representative nuclear and cytoplasmic stains for pAkt in normal appearing epithelium and lung tumors are presented in Figure 1. In normal airway epithelium, pAkt staining was nuclear, whereas the alveolar epithelial cells did not stain. We found that 83.3% (179/215) of NSCLCs exhibited positive nuclear staining for pAkt (index  2), whereas 16.7% (36/215) had strong pAkt nuclear staining (index 5) without histologic preference (Table 3). Positive cytoplasmic expression of pAkt (index 2) was similar to that of nuclear expression insofar as 87.9% (189/215) of tumors displayed a pAkt index 2 with only 12% (26/215) of tumors showing cytoplasmic pAkt levels at an index 5. The percentage of NSCLCs negative for both nuclear and cytoplasmic pAKT staining was 10.2% (22/215), 12.1% (26/215) of tumors were positive for cytoplasmic staining only, and 34.4% (74/215) were positive for nuclear staining only. In SCLCs, most samples tended to exhibit an intermediate level of staining (index 2–4) for both nuclear and cytoplasmic pAkt, 66.7% (8/12) and 50% (6/12), respectively. SCLCs with strong nuclear staining for pAkt (index 5) were limited to 8.3% of samples (1/12). A summary of the distribution of both nuclear and cytoplasmic pAkt staining in NSCLCs and SCLCs is presented in Table 3. PIK3CA gene copy number did not correlate with pAkt nuclear (p = 0.4904) or cytoplasmic (p = 0.6031) staining in lung cancer.
PIK3CA CN and pAkt Expression during Tumorigenesis To determine when PIK3CA amplification is observed during the course of tumor progression, we performed dual color FISH for PIK3CA in 43 preinvasive bronchial lesions obtained at the time of bronchoscopy from 39 patients. Clinical information was available on 14 patients, and showed that preinvasive lesions were synchronous to invasive tumors in all 14. We found an increased CN ratio for PIK3CA/FHIT in high-grade lesions (moderate dysplasia, severe dysplasia, carcinoma in situ [CIS]) that was significantly different from the CN ratio in low-grade lesions or in normal epithelium (ANOVA, p < 0.001) (Figure 2). Severe dysplasia and CIS were almost invariably amplified for PIK3CA (Figure 3).
We found weak immunostaining for pAkt in 8/10 normal bronchial specimens; the remaining two were negative. Whereas the bronchial smooth muscle, stroma, and alveolar tissue were negative, the submucosal glandular epithelium showed 2 + staining. In low-grade preinvasive lesions, immunostaining was weak, with only 2/12 (16.7%) showing focal moderate to strong immunopositivity. In contrast, in high-grade preinvasive lesions 11/16 (69%) showed moderate to strong immunopositivity throughout the thickness of the epithelium. Representative immunostains of preinvasive lesions are shown in Figure 4. There was a significant correlation between CN increase among the three groups (normal, low-grade, and high-grade lesions) and their respective immunostaining scores (Spearman correlation coefficient r = 0.4, p = 0.02).
In this study we demonstrate the frequent and early genomic amplification of PIK3CA in lung cancer. Downstream Akt was phosphorylated in a large majority of invasive and preinvasive lung cancers. The level of PIK3CA amplification did not correlate with pAkt immunostaining and is not associated with prognosis. PIK3CA genomic amplification and overexpression was found at the stage of severe dysplasia and likely represents a signature of commitment to tumor progression. These results further implicate the PI3-kinase/pAkt pathway in lung cancer development. Chromosome 3q26-ter amplification, including the PIK3CA gene locus at the peak of the amplicon, is one of the most prevalent genomic abnormalities in solid tumors and is likely to play a critical role in tumorigenesis. Amplification of chromosome 3q has been described in squamous epithelial transformation from the aerodigestive tract (19), bladder (20), cervix (21), and stomach (22). The 3q amplicon also contains a number of other important genes including p63, hTR, SKL, NEP, and somatostatin. The functions of some of these genes in NSCLC are being investigated. PIK3CA/Cent 3 ratio was increased in 43% (57/132) of lung cancers, which indicates a high frequency of amplification, particularly when compared with other common biomarkers: EGFR is amplified in 22% of NSCLCs (23); ras is mutated in 20–30% of lung cancers (24, 25); CCND1 (cyclin D1) is amplified in 8–30% of lung cancers (26, 27). Currently, these three genetic alterations and molecular pathways are being targeted for the treatment of lung cancer. In addition, the high prevalence of PIK3CA genomic amplification in preinvasive lesions (85%) and in squamous carcinoma (70%) makes this biomarker very attractive for the development of novel early detection strategies. Here we present further evidence implicating the PI3-kinase pathway in the development of lung cancer. Aberrations involving genes in the PI3-kinase pathway such as PIK3CA, PKB (or Akt), and PTEN have been implicated in tumorigenesis. After phosphorylation of Serine 473, pAkt exerts antiapoptotic properties through phosphorylation of BAD or caspase 9 (28), regulates the cell cycle by inhibiting cyclin D1 (29), and activates genes such as IkB (30). Akt or PKB is a major downstream target of receptor tyrosine kinases (such as EFGR, PDGFR) that signal through PI3-kinase and are responsible for many biological responses. We found very prevalent overexpression of pAkt in lung cancer without histologic preference. Our data are in concordance with the data presented by Lin and coworkers showing that overexpression of PIK3CA did not correlate with Akt2 upregulation (31). Specifically, there was no correlation between PIK3CA gene copy number and pAkt immunostaining. This observation points to the possibility of distinct PI3-kinase–dependent and –independent mechanisms of activation in NSCLC. In particular, PIK3CA amplification may not lead to phosphorylation of Akt, insofar as it is only one of the many downstream targets of PI3-kinase. In addition, phosphorylation of Akt was not found to be a prognostic marker for lung cancer. When we applied pAkt immunostaining to a large set of tumors, the results failed to show influence on survival in any histologic subtype. This suggests that pAkt may have a role in tumor development but not necessarily in progression to a metastatic phenotype. Understanding the mechanisms of Akt activation is important in fully exploring its usefulness as a therapeutic target. Currently, multiple pathways have been implicated as having roles in Akt activation. First, genomic amplification of PIK3CA or other genes in the pathway may induce changes in the regulation of Akt (5). Second, k-ras activation (32), growth factor activation, and oncogene activation (EGF, PDGF and IGF), have been shown to be effective mediators of Akt function (33). Third, PTEN is a phosphatase and a tumor suppressor gene that converts phosphatidylinositol-3-phosphate into PI-2 phosphate. Loss of PTEN is associated with an increase in Akt activity; however, its loss is only found in less than 10% of lung cancers (34). Finally, a recent report suggests that PIK3CA might be activated through a point mutation quite uncommon in lung cancer (35). Although these data would indicate that multiple pathways might be responsible for early and frequent Akt activation in lung cancer, inhibition of the PI3-kinase/Akt molecular pathway may have the most important therapeutic implications. For example, the identification of patients with increased pAkt expression may be useful in chemopreventive strategies. However, due to the wide range of intracellular functions possessed by PI3-kinase, a broad-spectrum antagonist to its kinase activity may have severe toxicity. Thus, more specific antiproliferative and proapoptotic PI3-kinase inhibitors are being actively pursued, particularly specific inhibitors of Akt (7). We attempted to validate pAkt immunostaining results by quantifying pAkt expression by Western analysis in a subset of 12 lung tumors. We were unable to demonstrate increased expression of pAkt as compared with total Akt. For unclear reasons, no one has been able to demonstrate pAkt activation in lung primary tumors, which is in contrast to results obtained in lung cell lysates (10). The nature of the antibody and the heterogeneity of lung tumors are likely explanations. Localization of pAkt in the nucleus and in the cytoplasm confirms previous reports (12, 36). In this study we attempted to obtain clinical information based on the localization of the staining without success. We know that PI3-kinase activity is required for translocation of Akt to the plasma membrane where phosphorylation occurs (37). Akt then phosphorylates a series of important proteins such as p21 and MDM2 in the cytoplasm to allow their nuclear localization (38, 39). Nonetheless, the precise role of Akt in the nucleus remains unclear. Because PIK3CA amplification appears to be prevalent in preinvasive lung cancer, this cytogenetic abnormality may be a valuable biomarker for identifying cancer early in patients with or at risk for lung cancer. Taking advantage of the frequent 3p deletion on 3p14 in NSCLC (40, 41), we assessed PIK3CA/FHIT CN ratio, increasing the sensitivity of our assay to detect small genomic abnormalities in tissue sections. Among preinvasive lesions, we found amplification of PIK3CA in high-grade lesions only. Low-grade preinvasive lesions found in the vicinity of invasive tumors did not show PIK3CA amplification. Given the quantifiable nature of PIK3CA CN by FISH and of pAkt by immunostaining, and to address obvious limitations of a retrospective study, a prospective evaluation of PIK3CA/pAkt as a predictive biomarker for lung cancer development in biopsy specimens from patients at risk for or with lung cancer is warranted. Taking into consideration the fact that not all high-grade lesions develop into invasive tumors (42), it will be critical to identify a population of high-risk individuals who do and do not develop lung cancer to validate these biomarkers as predictors of lung cancer development. The impact of this detection strategy relies on the ultimate ability to shift lung cancer diagnosis toward an earlier stage. This shift by itself could have a major impact on lung cancer management. Presently, there are a number of strategies being investigated for the purpose of achieving a transition toward early detection (sputum cytology, computed tomography), however, the use of biological techniques such as aCGH, FISH, and/or immunohistochemistry may be the most auspicious (43). Furthermore, PIK3CA and possibly pAkt may assist in monitoring the response to new chemopreventive agents. In conclusion, we provide further evidence of the frequent and early involvement of the PI3-kinase/Akt pathway in lung cancer. The role of the PI3-kinase/Akt pathway may have important implications in tumorigenesis, and genomic amplification of PIK3CA and pAkt overexpression may represent important biomarkers of tumor development. Consequently, further assessment of the PI3-kinase/Akt pathway should prove a worthwhile means of identifying genes and surrogate genetic endpoints useful in chemopreventive strategies.
The authors thank Candace Murphy and Darienne Adkins from the Lung SPORE Vanderbilt for their assistance in tissue array preparation and clinical research database development. They thank Drs. Kirk Jones and David Jablons for providing preinvasive tumor tissue sections from UCSF.
Supported in part by a Merit Review Entry Program grant from the Veterans Administration and a career development award from the SPORE in Lung Cancer 5P50 CA 90949–02. This article has an online data supplement, which is accessible from this issue's table of contents at www.atsjournals.org Conflict of Interest Statement: P.P.M. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; P.M.T. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; Y.S. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; S.M.J.R. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; P.Y. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; B.S. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; M.E.E. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; M.N. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; J.J.A. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; A.L.G. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form April 10, 2004; accepted in final form August 18, 2004
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
