Proteomic Strategies for the Characterization and the Early Detection of Lung Cancer

STATE

OF THE

ART: CONCISE REVIEW

Proteomic Strategies for the Characterization and the Early Detection of Lung Cancer Pierre P. Massion, MD,* and Richard M. Caprioli, PhD†

Abstract: Biochemical techniques focused on changes in the proteome in benign and malignant conditions have been applied to the study of lung cancer. Although relatively little information is currently available about the human lung proteome, the major goals of the analysis of lung cancer are to better understand tumor biology, to define early detection biomarkers and predictors of tumor behavior, and to identify potential new therapeutic targets. In this review, we summarized the last 10 years of research in this area with emphasis on its application to the early detection of lung cancer. Basic analytical tools, such as two-dimensional gel electrophoresis, mass spectrometry, and protein arrays, are described and placed in the context of lung cancer research. (J Thorac Oncol. 2006;1: 1027–1039)

P

roteins are responsible for the function of biological systems and the phenotypes of cells. Cancer cells express proteins that distinguish them from normal cells. In an effort to characterize the molecular determinants of lung cancer, it is critical that we probe these tissues and biological fluids with tools that address the biology of lung cancer directly at the protein level. Proteomics aims to characterize proteins to obtain a more integrated view of the biology. Proteomics attempts to answer how proteins are expressed, how they function and interact, how they get post-translationally modified after different injuries, and how they can participate as specific biomarkers of disease.1 As a result, new technologies are being developed to allow the rapid and systematic analysis of thousands of proteins. In particular, the identification of novel biomarkers to differentiate tumor from normal cells and predict individuals likely to develop lung cancer represents major clinical questions still largely unanswered. *Division of Allergy, Pulmonary and Critical Care Medicine and †Department of Biochemistry, Vanderbilt Ingram Cancer Center, Veterans Administration Medical Center, Nashville, Tennessee. This review was supported in part by CA102353 to PPM. PPM is a Damon Runyon-Lilly Clinical Investigator supported by the Damon Runyon Cancer Research Foundation (CI-#19-03). Address for correspondence: Pierre P. Massion, MD, Division of Allergy, Pulmonary and Critical Care Medicine, Vanderbilt-Ingram Comprehensive Cancer Center, 2220 Pierce Avenue, Preston Research Building 640, Nashville TN 37232-6838. E-mail: [email protected] Copyright © 2006 by the International Association for the Study of Lung Cancer ISSN: 1556-0864/06/0109-1027

Proteomics-based early detection strategies for cancer diagnosis include the analysis of complex mixtures such as tissue samples, serum, plasma, sputum, and exhaled breath condensate. The inherent analytical advantages of mass spectrometry (MS), including sensitivity and speed, promise to make MS a mainstay of biomarker discovery. The development of ionization techniques now allows almost any compound to be studied by MS. The optimal use of these technologies depends on the desired goal, such as protein identification, identification of post-translation modification, or determination of protein-protein interactions. Biomarkers of early detection of lung cancer are still at an early stage of development.2 The Early Detection Research Network (National Cancer Institute, division of cancer prevention) has proposed a step-wise method for evaluating biomarkers and identifying people at risk (http://www.cancer. gov/edrn).3 No current biomarkers for the early detection of lung cancer have passed the early validation (phase II).4 Whereas genetics has provided considerable insights into the molecular biology of lung cancer,5 the overall correlation between level of expression of the messenger RNA molecules and protein expression is relatively poor.6 Yet, we assume that biomarkers exist and that proteomic technologies offer a new avenue for biomarker discovery. The hypothesis behind protein-based early diagnosis of lung cancer is that a transformed cancer cell and its clonal expansion results in up- or down-regulation of specific hostor tumor-derived proteins, some of which will be secreted. These proteins should be detectable in biological specimens, including airway samples, sputum, and exhaled breath, or in other samples obtained by methods ranging from venopuncture, thoracentesis, bronchial brushing, lavage or biopsy, transthoracic needle aspirates, or surgical biopsy specimen. These proteins, including tumor antigens, can be detected by a variety of methods with different sensitivity and specificity. Historically, early detection has been most challenging because of the limited sensitivity of most tests applied to screening strategies, and a molecular approach may address some of the limitations of more traditional imaging or histological-based approaches. Recently, studies have shown that cancer-specific “fingerprints” can be obtained directly from tissue or serum and have great promise for the early diagnosis of cancer.7–10 In this review, we discuss the proteomic methodologies applied to lung cancer specimens, review the original published data, and propose future directions.

Journal of Thoracic Oncology • Volume 1, Number 9, November 2006

1027

Journal of Thoracic Oncology • Volume 1, Number 9, November 2006

Massion and Capriolo

METHODOLOGICAL APPROACHES From the traditional proteomic approaches discovering and assessing proteins one by one, the field has rapidly progressed in the last 10 years to high-throughput proteomic strategies. Identification and validation of biomarkers is accomplished by a) two-dimensional (2D) gel electrophoresis separation followed by analysis by time of flight mass spectrometry; b) proteomic profiling based on mass to charge intensities; c) a combination of liquid chromatography followed by MS in which proteolytic digestion and peptide separation allow identification by tandem mass spectrometry; or d) protein arrays (Table 1). When referring to the “top down” approach, a protein is separated from a complex mixture, purified, and identified by direct fragmentation in a mass spectrometer. Ideally, this approach has many advantages, but protein separation is difficult in practice. An alternative is the “bottom up” or “shotgun” approach, in which a complex mixture of proteins is first made more complex through enzymatic digestion, e.g., trypsin, followed by liquid chromatography tandem MS to identify the peptide fragments. A protein database search is then used to match the fragments (Figure 1). Because complementary information is often obtained, it is likely that a combination of these approaches will be necessary to for protein expression analysis.

Two-Dimensional Gel Electrophoresis Two-dimensional gel electrophoresis has been the primary method of comprehensive proteomic analysis of tissue samples. Proteins are separated on the basis of their charge through isoelectric focusing (first dimension) and are then

further separated in a polyacrylamide gel on the basis of their molecular weight (second dimension). When silver staining techniques are used, more than 1000 proteins can be visualized on a single gel. Stained gels are scanned with laser densitometers, and spot detection can be analyzed and quantified with software such as PDQUEST.11 Immobilized pH gradients,12 narrow range pH gradients,13and fluorescent dye labeling techniques14 have greatly improved the quality and sensitivity of the information obtained. Another development in 2D gels is the use of differential in-gel electrophoresis (DIGE), in which two pools of proteins are labeled with different fluorescent dyes.15 The labeled proteins are mixed together and analyzed in the same 2D gel. DIGE improves the reproducibility, sensitivity, and quantitative aspects of 2D gel analysis.16,17 Although 2D polyacrylamide gel electrophoresis (2D-PAGE) and related technologies have proven to be excellent tools for the discovery of proteins associated with pathological states and, when combined with MS, can lead to the identification of discriminating protein spots,18 only a small percentage of the proteome can be visualized by 2D-PAGE, and 2D gels remain a low-throughput approach that requires relatively large amounts of sample (important for clinically applicable proteomics).

Matrix-Assisted Laser Desorption Ionization Mass Spectrometry Development of ionization techniques such as electrospray ionization (ESI) and matrix-assisted laser-desorption ionization (MALDI), coupled with new analyzers such as

TABLE 1. Comparison of Proteomic Platforms for Biomarker Discovery 2D-PAGE

MALDI TOF MS

ESI-MS (LC-MS/MS)

Through-put

Low; labor intensive Increased with automation Low for low abundance proteins, proteins with extreme pI, molecular weights, and hydrophobicity

Moderate Single separations (RP) Low Multidimensional fractionations (10–15) High dependent on fractionations to reduce sample complexity

High Optimal

Sensitivity

Detection limits

10–100 kDa Detection dependent on stains used to visualize proteins

High Increased with automation (hundreds of samples/day) Low Typically high abundant proteins are favored. Low abundant proteins suppressed unless fractionation techniques applied Optimal range ⱕ50 kDa extended range 200 kDaResolution decreases at higher molecular weights

300–2000 Da. Higher molecular weight proteins are observed as multiply charged ions

Protein quantification

No but improved with 2D-DIGE technology using fluorescent dyes

No but improved with isotopic labeling analyses

Reproducibility Advantages

Good Automation, post-translational modifications Cumbersome to run, large amount of protein needed

No Isotope labeled interval standards allow targeted quantitation Good Rapid and small amount required Low % of proteins analyzed, no rapid identification

No Dependent on the sens. and spec. of the detection system. May require amplification of signal Yes

Limitations

1028

Protein Arrays

High dependent on the quality of the detection method (direct, antigen capture or sandwich immunosassays)

Poor Depth of the analysis

Good Automation, robustness

Needs fractionation Not compatible with all buffers

Requires prior identification of proteins, technical challenges

Copyright © 2006 by the International Association for the Study of Lung Cancer

Journal of Thoracic Oncology • Volume 1, Number 9, November 2006

Proteomics in Lung Cancer Research

FIGURE 1. Schematic representation of identification of the components of a complex protein mixture through top-down and bottom-up proteomic approaches. 2D-Page, two-dimensional polyacrylamide gel electrophoresis; MS-MS, tandem mass spectrometry; MALDI TOF MS; matrixassisted laser desorption time-of-flight mass spectrometry; LC MS-MS, liquid chromatography tandem mass spectrometry; Thermo LTQ, linear ion trap mass spectrometer.

TOF ion trap, and quadrupole technologies, have facilitated the characterization of proteins by MS. These ionization techniques transfer proteins into a gas phase as charged molecules, enabling analysis in the mass spectrometer. Digested peptide samples can be analyzed in a mass spectrometer through a microelectrospray,19 allowing for direct sequencing and identification. For example, a MALDI quadrupole TOF instrument (so-called TOF technology) that allows a combination of peptide mapping with peptide sequencing has been developed.20,21 This instrument allows identification of a protein either by peptide mass fingerprinting of the protein digest or from tandem mass spectra acquired by collision-induced dissociation of individual peptide precursors. A peptide mass map of the digest and tandem mass spectra of multiple peptide precursor ions can be acquired from the same sample in the course of a single experiment. In MALDI MS, analysis requires that the sample is crystallized with a matrix that absorbs energy and subsequently ejects and ionizes the molecules into the gas phase. A strong electrical field accelerates all the ions in a pulse so that ions reach the detector at a speed that is inversely proportional to their mass-to-charge (m/z) ratios (i.e., lighter ions arrive earlier and heavier ions later). This time-of-flight is then calibrated to m/z. Complex mixtures can be analyzed by MALDI MS without fractionation.22 This type of analysis allows the simultaneous determination of protein molecular masses from 1 to ⬎200 kDa with high accuracy. Two features of MALDI TOF MS that make it especially promising for MS analysis of biological samples are its capability for very high throughput, in which a sample can be analyzed in seconds, and its higher tolerance for salts, buffers, and other biological contaminants. Because of these qualities, MALDI MS has been used to study proteins/ peptides in serum, blood, urine, tissue extracts, whole cells, and laser-captured microdissected cells. MALDI MS has also

been successfully applied to cells from the mammary epithelium or colon crypt.23 MALDI MS imaging/profiling, depicted in Figure 2, is a new technology for direct mapping and high-resolution imaging of biomolecules present in tissue sections.24,25 In this system, frozen tissue sections or individual cells are mounted on a metal plate and spotted with droplets of matrix in a regular array spaced at 50 to 100 ␮m; mass spectra are acquired from each pixel; and the spectrum from each pixel is then queried for the presence of specific mass spectral peaks. If a small number of discrete areas are to be analyzed, then these are ablated individually in a profiling mode. If the relative intensity of any of the signals in the spectrum is to be measured over the entire tissue at higher image resolution, then a raster is performed to provide a matrix of pixels. An image representing the intensity of a given peak in each pixel in the array is then generated. Imaging MS shows potential for several applications, including biomarker discovery, biomarker tissue localization, understanding molecular complexity of tumor tissues, and assessment of surgical margins in resected tumors.24

Liquid Chromatography Tandem Mass Spectrometry Liquid chromatography (LC) MS-MS (Tandem MS) couples HPLC with MS and has had a great impact on profiling of small molecule and protein, on identification of protein and their posttranslational modifications, and has proven to be an important alternative method to 2D gels.26,27 Typically, a mixture of proteins is first digested with site-specific proteases, then the resulting peptides are separated by using LC, and fractions are then analyzed by tandem MS (MS/MS). In this procedure, a mixture of charged peptides is separated in the first MS according to their m/z ratios. Each of these ions, or only those desired, are then sequentially directed into a collision cell that fractures these peptides into fragment ions

Copyright © 2006 by the International Association for the Study of Lung Cancer

1029

Massion and Capriolo

Journal of Thoracic Oncology • Volume 1, Number 9, November 2006

the mass spectrometer can identify amino acids by measuring the precise differences in the molecular weights of these fragments, the sequence of the peptide can then be deduced from the resulting fragments. By comparing predicted sequences in the databases, the identity of the peptide and, therefore, the protein from which it came, is deduced. Electrospray ionization (ESI) is a technique by which peptides and proteins in solution are converted into multiply charged ions in a gas phase for analysis in a mass spectrometer. Basically, the ionization process uses small solvent droplets containing the molecule of interest. As the droplet dissolves in the process, the protons in the droplet interact with basic sites on the molecule, creating multiply protonated (multiply charged) ions. Thus, a protein of 50,000 MW may typically have 15 to 30 or more protons, bringing the m/z value of the molecule to approximately 2000, a range attainable by most MS instruments. These gas phase ions move to the analyzer under the influence of varying pressures and electric fields at the boundary leading to the mass analyzer.28 The multiple charges on the gas phase peptides bring the mass-to-charge ratio down to a few thousand from a molecular weight of tens of thousands. Because mass spectrometers directly measure mass per charge, this method allows the measurement of large molecules without exceeding the mass limit of the detector (usually approximately 2000 atomic mass units). ESI is usually coupled with a triple quadrupole, ion trap, or hybrid TOF (QTOF) MS. The ability to conduct tandem MS after ESI has revolutionized proteomics, because this MS/MS method can be used to directly obtain amino acid sequence information. ESI presents significant advantages in the ease of coupling compared with separation techniques such as LC and HPLC by allowing online analysis of protein or peptide mixtures. However, it is a relative low-throughput technology.

Protein Arrays

FIGURE 2. Schematic representation of matrix-assisted desorption ionization mass spectrometry. Fresh tissue is sectioned and mounted on a conductive metal plate. The sample is coated with matrix in a raster design, typically sinapinic acid, and directly analyzed in the mass spectrometer. The relative intensity of any of the signals in the spectrum is measured over the entire tissue at higher image resolution. If a small number of discrete areas are to be analyzed, then these are ablated individually in a profiling mode. If the relative intensity of any of the signals in the spectrum is to be measured over the entire tissue at higher image resolution, then a raster is performed to provide a matrix of pixels in an imaging mode. An image representing the intensity of a given peak in each pixel in the array is then generated.

derived from the “parent” species. Fragmentation occurs predominantly at the peptide bonds such that a ladder of fragments is generated. The fragment ions are separated according to their m/z in the second scanning MS. Because

1030

Protein microarray technology provides an in vitro means of studying function as reflected in protein alterations related to disease status. For this application, the proteins themselves are arrayed on a solid support. The protein spots may consist of antibodies, cell or phage lysates, or recombinant proteins or peptides.29 Detection of the array is achieved by probing with a tagged antibody, ligand, or serum/cell lysate. The signal generates a pattern of positive and negative spots. The signal intensity of each spot is proportional to the quantity of applied tagged molecules bound to the molecule. Significant difficulties in this technology relate to the detection of low abundance proteins with adequate sensitivity and specificity–the ability to page endogenous molecules such as peroxidases, biotin, or immunoglobulins, and that may be important because of their ability to interfere with the detection amplification chemistries. Autoantibodies can be particularly useful for studying cell-surface antigens on cancer cells and could become a powerful tool for screening large numbers of antigens by protein microarray.30 A reverse-phase protein array approach that immobilizes the whole repertoire of a tissue’s proteins has been developed.31 Despite its advantages, protein microarray technology has some limitations. This technology requires high-quality and comprehensive expression libraries and methods that

Copyright © 2006 by the International Association for the Study of Lung Cancer

Journal of Thoracic Oncology • Volume 1, Number 9, November 2006

allow the robust analysis of a large number of functionally active proteins. Lack of availability of high-affinity and high-specificity antibodies for gene products and post-translationally modified proteins is one of the major limitations of this technology.32

APPLICATION OF PROTEOMIC METHODOLOGIES TO LUNG CANCER AND RELATED BIOSPECIMENS Lung Tissue A comprehensive analysis of the proteome in lung cancer is being pursued.33 Hanash et al.33 have analyzed more than 1000 lung cancer-related samples using 2D PAGE, in combination with MS, and have constructed the lung proteomic database. The aim of their studies was to identify biomarkers for the early detection of cancer, for developing novel classification of tumors, and for revealing novel targets for therapeutic intervention (Figure 3).6,34 They performed parallel analysis of the transcriptome and of the proteome in lung

Proteomics in Lung Cancer Research

tumors, comparing mRNA and protein levels in the same tumors.6 The integrated intensities of 165 protein spots representing the protein products of 98 genes were analyzed in 76 lung adenocarcinomas and nine unaffected lung tissues using 2D PAGE. For the same 85 tissue samples, mRNA levels were determined using oligonucleotide microarrays. Only 21 of the 98 genes analyzed (21.4%) showed a statistically significant correlation between protein and mRNA levels. Glucose-regulated M(r) 58,000 protein stands out as a protein elevated at the mRNA and translational level. This does not mean that microarray studies are uninformative, rather that they convey additional information with only partial overlap with proteomic data. Recently, Beer et al.35 identified a battery of genes and related proteins validated using a training and independent testing set associated with survival of adenocarcinoma of the lung. Using 2D gel analysis, the same group identified five CK7 cleavage products best associated with patient survival, a subset of which correlated to gene expression.34 Using 2D

FIGURE 3. (A), 2D-PAGE gel separation of proteins identified with silver staining from a stage I lung adenocarcinoma. The proteins are separated by isoelectric point (PI) in the first dimension and by molecular weight (MW) in the second dimension. (B–F), the outlined areas of (A) showing proteins significantly increased in lung adenocarcinoma (Chen G, Gharib TG, Huang CC, et al. Proteomic analysis of lung adenocarcinoma: identification of a highly expressed set of proteins in tumors. Clin Cancer Res 2002;8:2298 –2305. Reprinted by permission. Copyright © 2006 by the International Association for the Study of Lung Cancer

1031

Massion and Capriolo

Journal of Thoracic Oncology • Volume 1, Number 9, November 2006

PAGE and MS, Chen et al. identified and characterized several lung cancer-specific protein markers of lung adenocarcinomas, such as antioxidant enzyme AOE372, ATP synthase subunit d(ATP5D), beta1,4- galactosyltransferase, cytosolic inorganic pyrophosphatase, glucose-regulated M(r) 58,000 protein, glutathione-S-transferase M4, prolyl 4- hydroxylase beta subunit, triosephosphate isomerase, and ubiquitin thiolesterase (UCHL1).6 They went on to identify 33 of 46 survival-associated proteins by MS. Expression of 12 candidate proteins was confirmed as tumor-derived with immunohistochemical analysis and tissue microarrays.36 Using a MALDI MS technology, Bergman et al.11 detected increases in the expression of cathepsin D in lung adenocarcinoma. Many of these proteins represent specific isoforms of known proteins and reflect post-translational modifications and potential degradation forms. They also confirmed these candidates by using both mRNA microarrays and tissue microarrays.35,36 None of those candidate biomarkers has a current application in the clinic.

After subcellular fractionation and substractive proteomics, investigators have identified endothelial cell surface proteins aminopeptidase-P and annexin A1, exhibiting restricted lung tissue distribution and over-expression in lung tumors. Radio-immunotherapy to annexin A1 destroyed tumors and increased animal survival.37 Most recently, a small series of surgically resected lung tumors were recently analyzed by 2D gel and MALDI MS to identify a series of proteins previously reported such as annexin II, cathepsin D, HSP27, stathmin, and MnSOD, confirming the validity of this methodology to identify candidate biomarkers.38 With recent advances in MS techniques, it is now possible to investigate protein expression profiles from small biological specimens over a wide range of molecular weights. Using this technology on tumor lysates, profiles of 10 nonsmall cell lung cancers were analyzed, and two proteins, MMIF and cyclophilin A, were identified as candidate biomarkers39 and later found not to be prognostic markers of disease based on a lung cancer tissue microarray (Figure 4).40

FIGURE 4. Tissue microarrays (TMAs) of lung cancer specimens. TMAs are comprised of core biopsies 0.6 mm in diameter of different tumors and of uninvolved lung from the same individuals. TMAs allow high-throughput analysis of molecular markers identified in lung cancer and control samples. Clinical annotation of these specimens allows rapid assessment of clinical outcomes (e.g., survival) as schematically represented by Kaplan-Meier survival curves. FISH, fluorescent in situ hybridization; IHC, immunohistochemistry.

1032

Copyright © 2006 by the International Association for the Study of Lung Cancer

Journal of Thoracic Oncology • Volume 1, Number 9, November 2006

Proteomics in Lung Cancer Research

FIGURE 5. Detection of the optimal discriminatory biomarker sets in lung tumors. (Top), Representative MALDI-TOF-MS spectra obtained from tumor and normal lung tissue samples with molecular weight calculation (m/z values). *Examples of the MS peaks identified by the statistical analyses as optimal discriminatory patterns between normal and tumor. (Bottom), Hierarchical cluster analysis of 42 lung tumors and eight normal lung tissues in the training cohort according to the protein expression patterns of 82 MS signals. Each row represents an individual proteomic signal, and each column represents an individual sample. The dendrogram at the top shows the similarity in protein expression profiles of the samples. Substantially raised (red) expression of the proteins is noted in individual tumor and normal lung tissue samples. AD, adenocarcinoma; SQ, squamous cell carcinoma; LA, large cell carcinoma; META, metastases to lung from other sites; REC, recurrent non-small cell lung cancer; CAR, pulmonary carcinoid; NL, normal lung. (Yanagisawa K, Shyr Y, Xu BJ, et al. Proteomic patterns of tumour subsets in non-small-cell lung cancer. Lancet 2003; 362:433– 439. Reprinted by permission.)

We and others have recently demonstrated that proteomic profiling of lung tumors using MALDI MS or 2D gel methods allows distinction between normal and cancer tissue, and it may predict lymph node involvement or survival.8,36,41 In a recent report, we used conventional MALDI MS to generate protein profiles that accurately classify and predict histological subgroups of lung cancer (Figure 5).8 Using biostatistical methods to select differentially expressed peaks (MS signals), and after the development of a class prediction model,42 82 discriminatory signals were found to classify normal lung from lung cancer tissue samples in a derivation and validation study design with excellent accuracy. Other recent studies indicated the importance of using protein

expression profiles as diagnostic or prognostic biomarkers for patients with early-stage lung cancer using a type of MALDI MS termed surface-enhanced laser desorption/ionization (SELDI) technology.43 Although preliminary, these studies stress the relevance of protein signatures for the diagnosis and evaluation of response to therapy in lung cancer. Functional proteomics is emerging as a powerful tool that uses proteomic technologies to investigate the interaction, integration, and functions of proteins. This strategy couples proteomic information with biochemical and physiological analysis to advance our understanding of the functional role of proteins in normal and diseased organs. The proteomic profiling of protease activity in cancer is an attrac-

Copyright © 2006 by the International Association for the Study of Lung Cancer

1033

Massion and Capriolo

Journal of Thoracic Oncology • Volume 1, Number 9, November 2006

tive example of this approach. Because proteases are involved in many aspects of tumor development, assaying their nature and post-translational modifications is of great interest. Activity-based proteomic profiling allows such screening with molecular tags specific for enzymatic activity.44 In particular, this approach can be associated with multidimensional proteomic LC-MS analysis to discover the relative roles of active proteases in tumor cells and its microenvironment.45

Serum and Plasma Analysis of the serum proteome assumes that tissue perfusion of tumors or host responses contribute to the modification of circulating protein or peptide concentrations. Therefore, the proteins present in the serum are hypothesized to reflect the pathological state of the organism. Thus, investigators attempt to uncover a discriminating pattern in a small subset of proteins (among thousands) that are useful for diagnostic purposes. Evaluation of the lower molecular weight protein profile can be accurately analyzed by using a variety of methods, including ESI/MS/MS and MALDI MS, and may correlate with these pathological states. It remains to be seen whether there is sufficient information in this slice of the blood proteome. Although they can be very challenging, serum proteomic studies have great promise for the early detection of cancer, the monitoring of disease status, the discovery of new targets for therapy, and the assessment of response to therapy. This new approach faces the difficulties of lack of standardization, continuously evolving technologies, limitations of bioinformatics tools, the complex nature and large dynamic range of the blood proteome, assay reproducibility, and identification and quantitation of candidate biomarkers. Proteomic studies using blood as a source of proteins still have not resolved the question of whether analysis of serum or plasma is more informative.46 Plasma requires the addition of an anticoagulant and, eventually, a two-step centrifugation to get rid of platelets. Clotting depletes fibrinogen from the serum but also depletes other proteins and is associated with the appearance of many peptides, likely cleavage fragments of circulating proteins. A systematic, rigorous, and prospective collection protocol for both plasma and serum is essential to answering these critical questions. The Human Plasma Proteome Project is an international collaboration to characterize proteins in the human serum and plasma. This organization is addressing in detail issues of protein identification; specimen collection, handling, and storage; depletion of highly abundant proteins; fractionation; and MS instruments for the identification of peptides from digested proteins. They have already identified more than 9500 proteins, 3020 of which have at least two peptide matches.47 Another argument revolves around the need for fractionation of the serum/plasma. The simple fact that 22 of these tens of thousands of proteins in serum constitute 99% of the total protein content in the serum and that there is a large concentration range of these proteins (approximately 108 fold) results in great difficulties in fractionating the serum and in identifying tumor-specific markers, many of which are likely to be present in low abundance. Fractionation is intu-

1034

itively more satisfying, reducing the complexity of the biological mixture, increasing the number of peaks obtained by MALDI MS, and increasing the likelihood of finding new biomarkers among lower abundant proteins. However, there are also disadvantages because fractionation requires larger sample amounts, is more time- and cost-consuming, and dramatically increases the risk of variability within and between samples, which may preclude the use of the resulting proteomic patterns for diagnosis. A dual approach is probably worth emphasizing at the early stage of development of this proteomic strategy. Recently published studies have reported serum protein expression profiles that distinguish cancer patients with a variety of malignancies from controls using various MSrelated approaches.9,10,48 –52 In the latter publication, the authors identified a serum protein pattern among patients with head and neck cancer who achieved a sensitivity of between 34% and 52% when applied to the serum of patients with lung cancer. Protein patterns were distinguished by using MALDI TOF MS combined with a simple classification procedure, based on a t test and linear discriminant analysis in tumor sera that included patients with head and neck or lung cancer.10 By using the optimal head and neck cancer model cutoff of 73% sensitivity and 90% specificity, they were able to discriminate squamous cell lung cancer with sensitivities of 52% (34% for adenocarcinoma and 40% for large cell carcinoma).10 However, this study was not designed to address whether this protein profile can discriminate patients with lung cancer from controls. This is a unique study that generates proteomic profiles in serum by using MALDI TOF MS, a very simple, rapid, and inexpensive technology that requires only basic sample processing and shows potential to achieve practical early cancer detection in biological fluids. A single protein by itself may not be useful for early detection, but it is thought that the most informative biomarker may be a combination of markers into signature protein profiles for early diagnosis. These studies are very encouraging and demonstrate that MALDI MS fingerprint protein profiling, coupled with a learning algorithm, can be useful for the early diagnosis of prostate, ovarian, breast, bladder, and aerodigestive cancers by using only 1 ␮L of serum. Our group recently generated protein expression profiles using MALDI TOF MS directly from 1 ␮L of unfractionated serum to define a discriminatory protein fingerprint to distinguish patients with lung cancer from matched controls.53 We found several proteins that could discriminate patients with early-stage lung cancer from controls; most of the proteins were in lower m/z ranges. Recently, automated tools for the discovery of serum markers have been developed. These apply small-volume robotics to serum fractionation and MALDI MS analysis. This approach may provide better reproducibility, multidimensionality, and high throughput to the analysis of biological specimens.54 This approach must be validated in larger populations and from a number of institutions before the further translation of this tool into general clinical decision-making for the early diagnosis of lung cancer. Shotgun proteomics is being investigated as a mode of

Copyright © 2006 by the International Association for the Study of Lung Cancer

Journal of Thoracic Oncology • Volume 1, Number 9, November 2006

in-depth analysis of the serum proteome for the purpose of biomarker discovery55 and as potential platform to be translating to the clinic. Questions revolving around reproducibility and rapidly changing technology are being considered. The search for autoantibodies as source of biomarkers is an attractive approach, as we know the sensitivity of the immune system to respond to increasing concentration of the peptide presented to the immune surveillance system. Authors have long searched for specific autoantibodies of cancer-related peptides (e.g., p53).56 More recently, investigators have proposed high-throughput screening of autoantibodies to circulating tumor antigens in patients with lung cancer.57 Briefly, 2D PAGE of the lung adenocarcinoma cell line (A549) was incubated with sera from patients with lung adenocarcinoma, other cancers, and no cancer controls. These studies found autoantibodies against glycosylated annexins I and II in 60% of the patients with lung adenocarcinomas and 33% of patients with squamous cell carcinomas and that this was associated with high circulating levels of IL-6.58 Similarly, studies involving sera from 64 newly diagnosed patients with lung cancer, 99 patients with other types of cancer, and 71 non-cancer controls revealed a tumor antigen, protein gene product 9.5 (PGP 9.5, a neurospecific protein) that induces a humoral response in lung cancer.57 In an effort to speed up the analysis of serum for autoantibodies to lungs bearing antigens, a novel approach that combines liquid phase protein separation with microarray technologies was developed.59 Whole cell or tissue lysates are fractionated by isoelectric focusing and reverse phase chromatography into hundreds of fractions, which are then arrayed onto nitrocellulose-coated slides and probed with antibodies against specific proteins. Such biochips require low sample volume and provide a rapid procedure to molecularly profile the antibody response to tumor antigens in cancer. Circulating antibodies to tumor-associated proteins were also isolated from the serum of patients with lung cancer from cDNA T7 phage libraries, and 45 candidates were identified. This is done by sequencing cDNA from phages of interest. Protein expression in serum is being confirmed by ELISA.60,61

Pleural Effusions Proteomics has the advantage of allowing the analysis of both tissue and biological fluids and the identification of proteins and protein post-translational modifications such as phosphorylation, glycosylation, and sulfation that could be specific for different tumor types. Nilsson et al. analyzed pleural exudates by applying 2D gel electrophoresis, MALDI TOF MS, and Western blotting.62,63 They confirmed the identity of proteins of potential diagnostic value such as cystatin C. Although it is possible to obtain mass spectra of proteins in complex biological mixtures (e.g., pleural effusion, cerebrospinal fluid) without previous purification steps,23,64 simple fractionation often results in a dramatic improvement of the signal-to-noise ratio. Because larger volumes of low concentrations of proteins can be loaded onto analytical 2D gels, low-abundance proteins in biological fluids, such as pleural exudates, may be enriched for further characterization by MS.

Proteomics in Lung Cancer Research

Recently, Bard et al. studied the presence of exosomes (membrane vesicles from endosomes) in cancerous pleural effusions and to identify their proteomic content.65 They found the presence of antigen-presenting molecules, cytoskeletal proteins, and signal transduction-involved proteins as well as SNX25, BTG1, PEDF, and thrombospondin 2. A high concentration of antigen-antibody humoral immunity components in exosomes may be important as an antigen source for cancer immunotherapy. Promising biomarkers of mesothelioma (i.e., mesothelin and osteopontin) are being introduced into clinical practice.66 – 68 Both were discovered by proteomic analysis and moved to phase 3 biomarkers validation. Osteopontin shows promise in distinguishing individuals with exposure to asbestos who do not have cancer from those with mesothelioma.

Exhaled Breath and Sputum Condensation of exhaled breath is a non-invasive way to collect material originating from the lung, including the lower respiratory tract. To date, exhaled breath condensate samples have been mainly studied for NO metabolites,69,70 8-isoprostane,71 and various inflammatory cytokines.72 EBC and saliva proteins have been characterized by using 2D electrophoresis,73 but characterization of the methods and the specimens has not yet been completed. Examining a set of volatile markers may enable recognition and diagnosis of diseases such as lung cancer. In the hope of evaluating the field of cancerization, the EBC has recently been studied in patients with lung cancer. Tumor suppressor gene p53 mutation was detected by PCR in exhaled breath condensate of patients with lung cancer.74 Endothelin-1 and microsatellite markers were found in EBC and therefore confirm the potential diagnostic value of this biological specimen.75,76 Although very appealing, there are technical challenges related to sampling and analysis, lack of normalization, and standardization; therefore, huge variations exist between results of different studies.5 Translating the analysis of EBC to clinical practice is likely to be problematic. The sputum is more difficult to analyze by proteomic methods than exhaled breath condensate. Traditional proteomics has used immunostaining of sputum samples for the validation of tumor-related antigens. Some of these markers are still being investigated in prospective trials.77 2D electrophoresis of sputum induction specimen is attractive, although specific alterations in protein composition in lung disorders are not well characterized. The soluble and cellular components may represent useful and separate components to analyze using MALDI MS. This work must be developed in the future.

Bronchial Biopsies In preliminary studies we used MALDI MS to identify specific patterns of protein expression of the airway epithelium in different histological stages of tumor progression. MALDI MS data were acquired from normal alveolar epithelium, normal bronchial epithelium, and preinvasive and invasive lung cancer tissues mainly from patients with concomitant lung cancer. Statistical analysis and supervised hierarchical cluster analysis revealed that protein profiles are

Copyright © 2006 by the International Association for the Study of Lung Cancer

1035

Massion and Capriolo

Journal of Thoracic Oncology • Volume 1, Number 9, November 2006

able to distinguish among groups with excellent accuracy.41 Based on this cross-section analysis of bronchial biopsies, we were able to predict the nature of the histological lesion. We hope to validate these data in an independent set of tissues and select proteins expressed in invasive and preinvasive, but not histologically normal epithelium. Ultimately, these proteins may provide us with targets for diagnostic markers and therapeutic intervention.

Bronchioalveolar Lavage Bronchioalveolar lavage (BAL) performed during fiberoptic bronchoscopy is a relatively safe technique that allows the collection of cells and a wide variety of soluble components, including proteins, from the human lung.78 Because the proteins present in BAL have many origins, and because of the diversity of proteins considered, analysis of the protein content of BAL is of great potential interest in defining biomarkers in many lung diseases. The potential for discovery of new lung disease markers in BAL via proteomics approaches is significant. The first 2D map displaying the major soluble proteins present in lung lavage was published in 197979 and pattern-matched with 2D maps of serum samples.80 After many different technical improvements, work still aims at the construction of a 2D reference database of BAL proteins and the understanding of the molecular processes involved in lung disease.81 Recently, the identification of proteins present in the BAL 2D map has been published.82– 84 The most important technical advancement that enabled BAL proteomic analysis was the improvement in the method used for isoelectric focusing. In 1990, Lenz et al.85 used improved methods for isoelectric focusing for the first-dimensional separation of proteins from dog BAL. The BAL sample preparation (salt removal, sample concentration, fractionation), sample loading technique, choice of pH range, and second-dimensional gel used are critical in obtaining optimal resolution of the widest range of proteins present in BAL.84,86 Several proteins related to lung inflammation have been identified; they are altered in smokers and are specific for different inflammatory conditions (sarcoidosis, interstitial pulmonary fibrosis, hypersensitivity pneumonitis). This BAL technique has not been studied in detail among patients at risk of or with lung cancer.

CHALLENGES AHEAD Proteomics studies currently use various methods for sample collection, preparation, storage, and processing, and for patient selection. The optimal steps in sample collection, processing, and analysis must be standardized.87,88 Because of the inherent instability of proteins, sample procurement and preservation issues must be addressed to make them compatible with proteomics analysis. Databases do not necessarily agree on criteria necessary to assign a protein identification. The Human Proteome Organization Plasma Proteome Project has more than 9600 proteins identified according to proposed internal criteria, but identification of proteins and statistical methods89 leading to the identification of proteins still are needed to reach agreement. The development of bioinformatics tools and optimization of biostatistical analysis are essential. The simplifica-

1036

tion of processing samples for a diagnostic test is an important goal of clinical proteomics. Obviously, the size of the datasets obtained after fractionation leads to difficulties in data analysis. The complex nature of the sample fractionation and data analysis also call for a major effort in mobilization of a process otherwise subject to relatively high variability. The integration of genomic with proteomic data represents a great challenge for the future. Many proteins that are expressed in different types of lung cancer have been identified and have been correlated with the expression measures for their corresponding genes at the RNA level. Because of the non-linear relationship among gene expression, translation to protein, and changes in function, it is important to assay for not only protein over-expression, but also their functional post-translational modifications. When analyzing biological samples, it is desirable to avoid excessive purification steps because loss of material is unavoidable at each step. The analysis of biological samples is also complicated by the fact that the sample contains many proteins, many in a low concentration. The scarcity of proteins of interest in biological materials makes isolation of sufficient material for proteomic analysis problematic, particularly for 2D electrophoresis. The peptidomic analysis of albumin-bound peptides may have some advantages.90,91 The more we fractionate proteins, the more we risk increasing variability. Quantitation of molecular species, particularly candidate biomarkers, is of critical importance. It is hoped that standard chemical assays, ELISA assays, and MS analysis will render the assays sensitive and reliable. Fluorescencebased and MS-based methodologies are being actively pursued: DIGE, multiplexed proteomics,92,93 isotope-coded affinity tagging (ICAT)15,94 and targeted quantitative analysis of proteins by isotopic dilution with labeled peptide internal standards (AQUA method)95 are being developed to this end.

CONCLUSIONS Integration of genomics and proteomics approaches will allow us to move closer to the goals of earlier detection. The rapid development of proteomic and genomic technologies has provided a large amount of novel information, has allowed comprehensive analysis of the molecular basis of disease, and is leading to the assembly of large protein inventories. Molecular profiling may assist in identifying high-risk populations, and it offers a unique opportunity to study early carcinogenesis and potentially to reduce cancer mortality through the available effective treatment modalities amenable to early cancer. Proteomic analysis has the potential to profile differences between lung tumor and no tumor, among different stages and histology of cancer, and among different cancer samples at the same stage of progression. The ability to identify important proteins involved in the transformation process may lead to early markers for the detection of specific types of cancers and treatments based on the molecular profile of lung cancer. In the past decade, lung cancer treatment has become increasingly target-oriented. However, all targets have been

Copyright © 2006 by the International Association for the Study of Lung Cancer

Journal of Thoracic Oncology • Volume 1, Number 9, November 2006

identified in late invasive and metastatic disease, thereby limiting the success of treatment. It is widely believed that if specific targets important in the earliest stage of the tumor can be identified, then application of these treatments early in the course of disease or premalignancy is likely to be more successful than attempting to treat late-stage tumors. Screening methods with high sensitivity and high specificity for early-stage lung cancer and that are also noninvasive, affordable, and safe, are urgently needed. Although measurements of chromosomal aberrations, point mutations, and loss of heterozygosity can be obtained from a series of biological specimens, the integration of these data with proteomic analysis will be critical in creating better clinical tools for the early detection and management of lung cancer. Clinical proteomics will have a fundamental impact on our understanding of complex disease processes, such as lung cancer, and will offer new opportunities in the diagnosis, prognosis, and therapy of disease. The development of specific and sensitive diagnostic biomarkers using biological fluids, such as sputum and serum, should improve screening, early detection, monitoring of disease progression, treatment response, and surveillance for recurrence.

17.

18.

19.

20.

21.

22. 23.

24.

25.

REFERENCES 1. Graves PR, Haystead TA. Molecular biologist’s guide to proteomics. Microbiol Mol Biol Rev 2002;66:39–63; table of contents. 2. Wardwell NR, Massion PP. Novel strategies for the early detection and prevention of lung cancer. Semin Oncol 2005;32:259–268. 3. Sullivan Pepe M, Etzioni R, Feng Z, et al. Phases of biomarker development for early detection of cancer. J Natl Cancer Inst 2001;93:1054–1061. 4. Network, T. E. D. R. Translational research to identify early cancer and cancer risk. Frederick, MD: NCI, Division of Cancer Prevention. 2005. 5. Chanin TD, Merrick DT, Franklin WA, Hirsch FR. Recent developments in biomarkers for the early detection of lung cancer: perspectives based on publications 2003 to present. Curr Opin Pulm Med 2004;10: 242–247. 6. Chen G, Gharib TG, Huang CC, et al. Proteomic analysis of lung adenocarcinoma: identification of a highly expressed set of proteins in tumors. Clin Cancer Res 2002;8:2298–2305. 7. Caprioli RM. Deciphering protein molecular signatures in cancer tissues to aid in diagnosis, prognosis, and therapy. Cancer Res 2005;65:10642– 10645. 8. Yanagisawa K, Shyr Y, Xu BJ, et al. Proteomic patterns of tumour subsets in non-small-cell lung cancer. Lancet 2003;362:433–439. 9. Adam BL, Qu Y, Davis JW, et al. Serum protein fingerprinting coupled with a pattern-matching algorithm distinguishes prostate cancer from benign prostate hyperplasia and healthy men. Cancer Res 2002;62: 3609–3614. 10. Sidransky D, Irizarry R, Califano JA, et al. Serum protein MALDI profiling to distinguish upper aerodigestive tract cancer patients from control subjects. J Natl Cancer Inst 2003;95:1711–1717. 11. Bergman AC, Benjamin T, Alaiya A, et al. Identification of gel-separated tumor marker proteins by mass spectrometry. Electrophoresis 2000;21:679– 686. 12. Gorg A, Obermaier C, Boguth G, et al. The current state of twodimensional electrophoresis with immobilized pH gradients. Electrophoresis 2000;21:1037–1053. 13. Tonella L, Walsh BJ, Sanchez JC, et al. ’98 Escherichia coli SWISS2DPAGE database update. Electrophoresis 1998;19:1960–1971. 14. Chambers AF, Naumov GN, Vantyghem SA, Tuck AB. Molecular biology of breast cancer metastasis: clinical implications of experimental studies on metastatic inefficiency. Breast Cancer Res 2000;2:400–407. 15. Patton WF. Detection technologies in proteome analysis. J Chromatogr B Analyt Technol Biomed Life Sci 2002;771:3–31. 16. Unlu M, Morgan ME, Minden JS. Difference gel electrophoresis: a

26.

27.

28.

29. 30.

31.

32.

33. 34.

35.

36.

37.

38.

39.

Proteomics in Lung Cancer Research

single gel method for detecting changes in protein extracts. Electrophoresis 1997;18:2071–2077. Zhou G, Li H, DeCamp D, et al. 2D differential in-gel electrophoresis for the identification of esophageal scans cell cancer-specific protein markers. Mol Cell Proteomics 2002;1:117–124. Li G, Waltham M, Anderson NL, Unsworth E, Treston A, Weinstein JN. Rapid mass spectrometric identification of proteins from two-dimensional polyacrylamide gels after in gel proteolytic digestion. Electrophoresis 1997;18:391–402. Emmett MR, Andren PE, Caprioli RM. Specific molecular mass detection of endogenously released neuropeptides using in vivo microdialysis/ mass spectrometry. J Neurosci Methods 1995;62:141–147. Krutchinsky AN, Zhang W, Chait BT. Rapidly switchable matrixassisted laser desorption/ionization and electrospray quadrupole-timeof-flight mass spectrometry for protein identification. J Am Soc Mass Spectrom 2000;11:493–504. Merchant M, Weinberger SR. Recent advancements in surface-enhanced laser desorption/ionization-time of flight-mass spectrometry. Electrophoresis 2000;21:1164–1177. Beavis RC, Chait BT. Rapid, sensitive analysis of protein mixtures by mass spectrometry. Proc Natl Acad Sci USA 1990;87:6873–6877. Xu BJ, Caprioli RM, Sanders ME, Jensen RA. Direct analysis of laser capture microdissected cells by MALDI mass spectrometry. J Am Soc Mass Spectrom 2002;13:1292–1297. Chaurand P, Fouchecourt S, DaGue BB, et al. Profiling and imaging proteins in the mouse epididymis by imaging mass spectrometry. Proteomics 2003;3:2221–2239. Stoeckli M, Chaurand P, Hallahan DE, Caprioli RM. Imaging mass spectrometry: a new technology for the analysis of protein expression in mammalian tissues. Nat Med 2001;7:493–496. McCormack AL, Schieltz DM, Goode B, et al. Direct analysis and identification of proteins in mixtures by LC/MS/MS and database searching at the low-femtomole level. Anal Chem 1997;69:767–776. Peng J, Elias JE, Thoreen CC, Licklider LJ, Gygi SP. Evaluation of multidimensional chromatography coupled with tandem mass spectrometry (LC/LC-MS/MS) for large-scale protein analysis: the yeast proteome. J Proteomics Res 2003;2:43–50. Kebarle P. A brief overview of the present status of the mechanisms involved in electrospray mass spectrometry. J Mass Spectrom 2000;35: 804–817. MacBeath G, Schreiber SL. Printing proteins as microarrays for highthroughput function determination. Science 2000;289:1760–1763. Robinson WH, DiGennaro C, Hueber W, et al. Autoantigen microarrays for multiplex characterization of autoantibody responses. Nat Med 2002; 8:295–301. Paweletz CP, Charboneau L, Bichsel VE, et al. Reverse phase protein microarrays which capture disease progression show activation of prosurvival pathways at the cancer invasion front. Oncogene 2001;20:1981– 1989. Espina V, Mehta AI, Winters ME, et al. Protein microarrays: molecular profiling technologies for clinical specimens. Proteomics 2003;3:2091– 2100. Oh JM, Brichory F, Puravs E, et al. A database of protein expression in lung cancer. Proteomics 2001;1:1303–1319. Gharib TG, Chen G, Wang H, et al. Proteomic analysis of cytokeratin isoforms uncovers association with survival in lung adenocarcinoma. Neoplasia 2002;4:440–448. Beer DG, Kardia SL, Huang CC, et al. Gene-expression profiles predict survival of patients with lung adenocarcinoma. Nat Med 2002;8:816– 824. Chen G, Gharib TG, Wang H, et al. Protein profiles associated with survival in lung adenocarcinoma. Proc Natl Acad Sci USA 2003;100: 13537–13542. Oh P, Li Y, Yu J, et al. Subtractive proteomic mapping of the endothelial surface in lung and solid tumours for tissue-specific therapy. Nature 2004;429:629–635. Alfonso P, Catala M, Rico-Morales ML, et al. Proteomic analysis of lung biopsies: differential protein expression profile between peritumoral and tumoral tissue. Proteomics 2004;4:442–447. Campa MJ, Wang MZ, Howard B, Fitzgerald MC, Patz Jr. EF. Protein expression profiling identifies macrophage migration inhibitory factor

Copyright © 2006 by the International Association for the Study of Lung Cancer

1037

Massion and Capriolo

40.

41. 42.

43.

44.

45.

46.

47.

48.

49. 50.

51. 52.

53. 54.

55.

56.

57.

58.

59.

60.

61.

Journal of Thoracic Oncology • Volume 1, Number 9, November 2006

and cyclophilin A as potential molecular targets in non-small cell lung cancer. Cancer Res 2003;63:1652–1656. Howard BA, Wang MZ, Campa MJ, Corro C, Fitzgerald MC, Patz Jr. EF. Identification and validation of a potential lung cancer serum biomarker detected by matrix-assisted laser desorption/ionization-time of flight spectra analysis. Proteomics 2003; 3:1720–1724. Rahman SM, Shyr Y, Yildiz PB, et al. Proteomic patterns of preinvasive bronchial lesions. Am J Respir Crit Care Med 2005;172:1556–1562. Shyr Y, Kim K. Weighted flexible compound covariate method for classifying microarray data. In D Berrar (Ed.), A Practical Approach to Microarray Data Analysis. New York: Kluwer Academic, 2003:186– 200. Zhukov TA, Johanson RA, Cantor AB, Clark RA, Tockman MS. Discovery of distinct protein profiles specific for lung tumors and pre-malignant lung lesions by SELDI mass spectrometry. Lung Cancer 2003;40:267–279. Saghatelian A, Jessani N, Joseph A, Humphrey M, Cravatt BF. Activitybased probes for the proteomic profiling of metalloproteases. Proc Natl Acad Sci USA 2004;101:10000–10005. Jessani N, Humphrey M, McDonald WH, et al. Carcinoma and stromal enzyme activity profiles associated with breast tumor growth in vivo. Proc Natl Acad Sci USA 2004;101:13756–13761. Thadikkaran L, Siegenthaler MA, Crettaz D, Queloz PA, Schneider P, Tissot JD. Recent advances in blood-related proteomics. Proteomics 2005;5:3019–3034. Rai AJ, Stemmer PM, Zhang Z, et al. Analysis of Human Proteome Organization Plasma Proteome Project (HUPO PPP) reference specimens using surface enhanced laser desorption/ionization-time of flight (SELDI-TOF) mass spectrometry: multi-institution correlation of spectra and identification of biomarkers. Proteomics 2005;5:3467–3474. Li J, Zhang Z, Rosenzweig J, Wang YY, Chan DW. Proteomics and bioinformatics approaches for identification of serum biomarkers to detect breast cancer. Clin Chem 2002;48:1296–1304. Banez LL, Prasanna P, Sun L, et al. Diagnostic potential of serum proteomic patterns in prostate cancer. J Urol 2003;170(2 Pt 1):442–446. Mobley JA, Lam YW, Lau KM, et al. Monitoring the serological proteome: the latest modality in prostate cancer detection. J Urol 2004;172:331–337. Petricoin EF, Ardekani AM, Hitt BA, et al. Use of proteomic patterns in serum to identify ovarian cancer. Lancet 2002;359:572–577. Petricoin III EF Ornstein DK, Paweletz CP, et al. Serum proteomic patterns for detection of prostate cancer. J Natl Cancer Inst 2002;94: 1576–1578. Yildiz P, Shyr Y, Rahman S, et al. Diagnosis of lung cancer from serum protein profile analysis. AACR Proc 45:A3335 (abstract, 2004). Villanueva J, Philip J, Entenberg D, et al. Serum peptide profiling by magnetic particle-assisted, automated sample processing and MALDI-TOF mass spectrometry. Anal Chem 2004;76:1560–1570. Tang HY, Ali-Khan N, Echan LA, Levenkova N, Rux JJ, Speicher DW. A novel four-dimensional strategy combining protein and peptide separation methods enables detection of low-abundance proteins in human plasma and serum proteomes. Proteomics 2005;5:3329–3342. Winter SF, Sekido Y, Minna JD, et al. Antibodies against autologous tumor cell proteins in patients with small-cell lung cancer: association with improved survival. J Natl Cancer Inst 1993;85:2012–2018. Brichory F, Beer D, Le Naour F, Giordano T, Hanash S. Proteomicsbased identification of protein gene product 9.5 as a tumor antigen that induces a humoral immune response in lung cancer. Cancer Res 2001; 61:7908–7912. Brichory FM, Misek DE, Yim AM, et al. An immune response manifested by the common occurrence of annexins I and II autoantibodies and high circulating levels of IL-6 in lung cancer. Proc Natl Acad Sci USA 2001;98:9824–9829. Madoz-Gurpide J, Wang H, Misek DE, Brichory F, Hanash SM. Protein based microarrays: a tool for probing the proteome of cancer cells and tissues. Proteomics 2001;1:1279–1287. Zhong L, Hidalgo GE, Stromberg AJ, Khattar NH, Jett JR, Hirschowitz EA. Using protein microarray as a diagnostic assay for non-small cell lung cancer. Am J Respir Crit Care Med 2005;172:1308–1314. Zhong L, Peng X, Hidalgo GE, Doherty DE, Stromberg AJ, Hirschowitz EA. Identification of circulating antibodies to tumor-associated proteins

1038

62.

63.

64.

65.

66.

67. 68. 69.

70.

71.

72.

73. 74. 75.

76.

77. 78. 79. 80.

81. 82.

83. 84.

85.

86.

for combined use as markers of non-small cell lung cancer. Proteomics 2004;4:1216–1225. Nilsson CL, Brodin E, Ekman R. Substance P and related peptides in porcine cortex: whole tissue and nuclear localization. J Chromatogr A 1998;800:21–27. Nilsson CL, Puchades M, Westman A, Blennow K, Davidsson P. Identification of proteins in a human pleural exudate using two-dimensional preparative liquid-phase electrophoresis and matrix-assisted laser desorption/ionization mass spectrometry. Electrophoresis 1999;20:860– 865. Westman A, Nilsson CL, Ekman R. Matrix-assisted laser desorption/ ionization time-of-flight mass spectrometry analysis of proteins in human cerebrospinal fluid. Rapid Commun Mass Spectrom 1998;12:1092– 1098. Bard MP, Hegmans JP, Hemmes A, et al. Proteomic analysis of exosomes isolated from human malignant pleural effusions. Am J Respir Cell Mol Biol 2004;31:114–121. Dennis JL, Hvidsten TR, Wit EC, et al. Markers of adenocarcinoma characteristic of the site of origin: development of a diagnostic algorithm. Clin Cancer Res 2005;11:3766–3772. Pass HI, Lott D, Lonardo F. Asbestos exposure, pleural mesothelioma, and serum osteopontin levels. N Engl J Med 2005;353:1564–1573. Robinson BW, Creaney J, Nowak A, et al. Mesothelin-family proteins and diagnosis of mesothelioma. Lancet 2003;362:1612–1616. Balint B, Donnelly LE, Hanazawa T, Kharitonov SA, Barnes PJ. Increased nitric oxide metabolites in exhaled breath condensate after exposure to tobacco smoke. Thorax 2001;56:456–461. Corradi M, Montuschi P, Donnelly LE, Pesci A, Kharitonov SA, Barnes PJ. Increased nitrosothiols in exhaled breath condensate in inflammatory airway diseases. Am J Respir Crit Care Med 2001;163:854–858. Carpenter CT, Price PV, Christman BW. Exhaled breath condensate isoprostanes are elevated in patients with acute lung injury or ARDS. Chest 1998;114:1653–1659. Scheideler L, Manke HG, Schwulera U, Inacker O, Hammerle H. Detection of nonvolatile macromolecules in breath: a possible diagnostic tool? Am Rev Respir Dis 1993;148:778–784. Griese M, Noss J, Bredow C. Protein pattern of exhaled breath condensate and saliva. Proteomics 2002;2:690–696. Gessner C, Kuhn H, Pankau HJ, et al. Factors influencing breath condensate volume. Pneumologie 2001;55:414–419. Carpagnano GE, Foschino-Barbaro MP, Mule G, et al. 3p microsatellite alterations in exhaled breath condensate from patients with non-small cell lung cancer. Am J Respir Crit Care Med 2005;172:738–744. Carpagnano GE, Foschino-Barbaro MP, Resta O, Gramiccioni E, Carpagnano F. Endothelin-1 is increased in the breath condensate of patients with non-small-cell lung cancer. Oncology 2004;66:180–184. Mulshine JL, Cuttitta F, Tockman MS, De Luca LM. Lung cancer evolution to preinvasive management. Clin Chest Med 2002;23:37–48. Reynolds HY. Use of bronchoalveolar lavage in humans: past necessity and future imperative. Lung 2000;178:271–293. Bell DY, Hook GE. Pulmonary alveolar proteinosis: analysis of airway and alveolar proteins. Am Rev Respir Dis 1979;119:979–990. Bell DY, Haseman JA, Spock A, McLennan G, Hook GE. Plasma proteins of the bronchoalveolar surface of the lungs of smokers and nonsmokers. Am Rev Respir Dis 1981;124:72–79. Noel-Georis I, Bernard A, Falmagne P, Wattiez R. Database of bronchoalveolar lavage fluid proteins. J Chromatogr B 2002;771:221–236. Wattiez R, Hermans C, Bernard A, Lesur O, Falmagne P. Human bronchoalveolar lavage fluid: two-dimensional gel electrophoresis, amino acid microsequencing and identification of major proteins. Electrophoresis 1999;20:1634–1645. Wattiez R, Falmagne P. Proteomics of bronchoalveolar lavage fluid. J Chromatogr B 2005;815:169–178. Lindahl M, Stahlbom B, Tagesson C. Two-dimensional gel electrophoresis of nasal and bronchoalveolar lavage fluids after occupational exposure. Electrophoresis 1995;16:1199–1204. Lenz AG, Meyer B, Weber H, Maier K. Two-dimensional electrophoresis of dog bronchoalveolar lavage fluid proteins. Electrophoresis 1990; 11:510–513. Sabounchi-Schutt F, Astrom J, Eklund A, Grunewald J, Bjellqvist B.

Copyright © 2006 by the International Association for the Study of Lung Cancer

Journal of Thoracic Oncology • Volume 1, Number 9, November 2006

87. 88. 89. 90.

Detection and identification of human bronchoalveolar lavage proteins using narrow-range immobilized pH gradient DryStrip and the paper bridge sample application method. Electrophoresis 2001;22:1851–1860. Baggerly KA, Morris JS, Coombes KR. Reproducibility of SELDI-TOF protein patterns in serum: comparing data sets from different experiments. Bioinformatics 2004;20:777–785. Baggerly KA, Morris JS, Edmonson SR, Coombes KR. Signal in noise: evaluating reported reproducibility of serum proteomic tests for ovarian cancer. J Natl Cancer Inst 2005;97:307–309. Fenyo D, Beavis RC. A method for assessing the statistical significance of mass spectrometry-based protein identifications using general scoring schemes. Anal Chem 2003;75:768–774. Tirumalai RS, Chan KC, Prieto DA, Issaq HJ, Conrads TP, Veenstra TD. Characterization of the low molecular weight human serum proteome. Mol Cell Proteomics 2003;2:1096–1103.

Proteomics in Lung Cancer Research

91. Lowenthal MS, Mehta AI, Frogale K, et al. Analysis of albuminassociated peptides and proteins from ovarian cancer patients. Clin Chem 2005;51:1933–1945. 92. Templin MF, Stoll D, Bachmann J, Joos TO. Protein microarrays and multiplexed sandwich immunoassays: what beats the beads? Comb Chem High Throughput Screen 2004;7:223–229. 93. Steinberg TH, Agnew BJ, Gee KR, et al. Global quantitative phosphoprotein analysis using Multiplexed Proteomics technology. Proteomics 2003;3:1128–1144. 94. Gygi SP, Rist B, Griffin TJ, Eng J, Aebersold R. Proteome analysis of low-abundance proteins using multidimensional chromatography and isotope-coded affinity tags. J Proteome Res 2002;1:47–54. 95. Gerber SA, Rush J, Stemman O, Kirschner MW, Gygi SP. Absolute quantification of proteins and phosphoproteins from cell lysates by tandem MS. Proc Natl Acad Sci USA 2003;100:6940–6945.

Copyright © 2006 by the International Association for the Study of Lung Cancer

1039