Chest ACCP Member Benefits
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Guest Access | Sign In via User Name/Password
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (7)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bowler, R. P.
Right arrow Articles by Reisdorph, N.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bowler, R. P.
Right arrow Articles by Reisdorph, N.
Related Content
Right arrow Translating Basic Research into Clinical Practice
(Chest. 2006;130:567-574.)
© 2006 American College of Chest Physicians

Proteomics in Pulmonary Medicine*

Russell P. Bowler, MD, PhD; Misoo C. Ellison, PhD and Nichole Reisdorph, PhD

* From the Department of Medicine (Drs. Bowler and Ellison), National Jewish Medical and Research Center; and the Department of Anesthesia (Dr. Reisdorph), University of Colorado Health Sciences Center, Denver, CO.

Correspondence to: Russell P. Bowler, MD, PhD, National Jewish Medical and Research Center, K729a, 1400 Jackson St, Denver, CO 80206; e-mail: BowlerR{at}njc.org

Abstract

Proteomics is the study of the entire protein complement of the genome (the proteome) in a biological system. Proteomic studies require a multidisciplinary approach and have only been practical with the convergence of technical and methodologic improvements including the following: advances in mass spectrometry and genomic sequencing that now permit the identification and relative quantization of small amounts (femtomole) of nearly any single protein; new methods in gel electrophoresis that allow the detection of subtle changes in protein expression, including posttranslational modifications; automation and miniaturization that permit high-throughput analysis of clinical samples; and new bioinformatics and computational methods that facilitate analysis and interpretation of the abundant data that are generated by proteomics experiments. This convergence makes proteomics studies practical for pulmonary researchers using BAL fluid, lung tissue, blood, and exhaled breath condensates, and will facilitate the research of complex, multifactorial lung diseases such as acute lung injury and COPD. This review describes how proteomics experiments are conducted and interpreted, their limitations, and how proteomics has been used in clinical pulmonary medicine.

Key Words: bioinformatics • electrophoresis • mass spectrometry

Similar to genomics, there are > 27 different definitions of "proteomics," yet all encompass the central concept of studying a nearly comprehensive set of proteins (as opposed to genes) expressed by a cell or organism. There are multiple advantages to studying protein expression rather than gene expression. First, it is primarily proteins, not genes, that determine how a cell functions. Second, a single gene or even a single mature messenger RNA may be associated with multiple proteins due to splicing, RNA editing, or posttranslational modifications, and there may be little correlation between messenger RNA levels and protein expression. Third, acellular compartments such as plasma and lung epithelial lining fluid (ELF) have little DNA or RNA but may have abundant proteins that are important markers of disease. A major disadvantage to proteomics vs genomics is technical. Nearly all genes are represented by only four nucleic acids, which can easily be detected and quantified using complementary nucleic acid sequences, whereas proteins are represented by > 20 amino acids and hundreds of distinct posttranslational modifications that require complicated identification and quantification (see following).

These technical limitations have been overcome, and proteomics research and publications have grown exponentially, primarily due to advances in protein separation, mass spectrometry (MS), and bioinformatics. This review will summarize common proteomics methods, discuss their limitations, and illustrate how proteomics has been used to study pulmonary biology.

Proteomic Methods

A typical workflow for a proteomics experiments includes the following: (1) sample acquisition and storage, (2) sample preparation and fractionation, (3) protein quantification and identification, and (4) bioinformatics.

Sample Acquisition and Storage
Proper study design is the first step and has been written about in depth elsewhere.1 Sample accessibility is the reason that most clinical studies use blood draws. Other factors are dependent on methodological approaches and may not be specific to proteomic studies (eg, subsets of proteins have diurnal expression patterns, heparinized blood tubes preserve different proteins from citrated tubes, increased freeze-thaw cycles may lead to differential protein degradation). Problems specific to proteomics include the use of certain reagents (eg, polyethylene glycol) or contamination (eg, by keratins). Inattention to these details is a frequent criticism of proteomics experiments; for example, it has been speculated that inconsistent sample processing has led to false ovarian cancer biomarker discovery using surface-enhanced laser desorption ionization (SELDI).2 Thus, investigators should be cautious and include appropriate experimental controls because any sample manipulations prior to proteomic analysis could add notable variation to protein measurement.

Preparation and Fractionation of Proteins
A typical complex biological sample such as serum or lung tissue has many thousands of proteins in concentrations that span > 10 orders of magnitude. For instance, serum albumin has a normal concentration range of 35 to 50 mg/mL (35 to 50 x 109 pg/mL), yet interleukin 6 has a normal range of just 0 to 5 pg/mL. Since no current technologies are capable of simultaneously resolving this number of proteins at such a great magnitude of concentrations, additional steps are needed to simplify the complexity of the sample, including separation by size, pH, and chromatographic properties, or by immunodepletion and enrichment strategies.

Size Separation: The most common laboratory method for separating proteins by molecular weight is sodium dodecyl sulfate-polyacrylamide gel electrophoresis (PAGE). Size separation by sodium dodecyl sulfate-PAGE relies on the property that protein migration in a constant electromagnetic field is roughly proportional to the molecular weight of a protein. Other methods include size exclusion (eg, column and membrane) and differential centrifugation. However, some reagents used with differential centrifugation are not compatible with direct downstream MS methods (eg, cesium chloride).

pH Separation: All proteins are made of acid, basic, and neutral amino acids. By altering the hydrogen concentration (pH) of a protein solution, one can force acidic and basic amino acids to change charge. By creating a pH gradient within an electric field, proteins will migrate to the point at which they are neutrally charged (ie, the isoelectric point [pI] of the protein). Advances in the manufacture of immobilized ampholyte gradients have revolutionized this approach by creating reproducible gel strips that include the pI of nearly all proteins (pI 3–10). Separation by pH (first dimension) is typically followed by PAGE (second dimension) to obtain a two-dimensional electrophoresis (2-DE) separation map. Certain classes of proteins are not well-suited for 2-DE, including membrane and insoluble proteins, low abundant proteins, and proteins with high molecular mass or extreme pH. In addition, several compounds interfere with the first dimension including salts, ionic detergents, and lipids. A major problem to 2-DE experiments is gel-to-gel variations that make comparisons difficult. A solution to this problem has been to label proteins from different samples with small molecular weight, electrically neutral fluorescent dyes, and then combining the samples before 2-DE separation.3 This technique has been termed differential in-gel electrophoresis (DIGE) and is useful for comparing two to three samples, although the cost of dyes can make these studies cost-prohibitive. MS is typically used downstream of 2-DE and DIGE to identify proteins of interest. Protein stains vary in their compatibility with MS, although protocols are available for processing commonly used stains (eg, silver, Coomassie, and Sypro Ruby).

Chromatographic Separation: Chromatographic separation relies on a the inherent difference of a protein in affinities for chemical substances. Reverse-phase chromatography is the most frequently used technique for proteomics and refers to the principle that hydrophobic substances (peptides and proteins) will elute from hydrophobic columns (eg, C4, C8, and C18) at progressively higher concentrations of organic solvent. Column-based liquid chromatography (LC) can be directly coupled to a mass spectrometer. Alternatively, surface chromatography (ie, SELDI) can be used to selectively capture proteins to a chip surface quantified using a mass spectrometer. A limitation of SELDI is that it is not always easy to identify the proteins that make up the profiles.

Immunodepletion and Enrichment Strategies: Immunodepletion, immunoprecipitation, and ultracentrifugation are strategies that improve detection on nonabundant proteins. Immunodepletion strategies remove high-abundance proteins such as albumin and Ig in order to detect medium-abundance and low-abundance proteins. Enrichment strategies rely on the ability of a specific antibody to pull out a single protein or related proteins from a complex solution. Ultracentrifugation of plasma has been successfully used to identify the proteins of plasma microparticles, which are spherical cell membrane fragments derived from either apoptotic or activated cells.4

Protein Identification and Quantification Using MS
MS: MS is one of the most common technologies used in proteomics (see reviews567) and refers to an instrumental method for identifying the chemical constitution of a substance by means of the separation of gaseous ions according to their differing mass and charge (Fig 1 ). Mass spectrometers can be used for profiling both peptides and small and large proteins. In a typical protein identification workflow, a protein typically is first digested using a proteolytic enzyme (eg, trypsin) that cleaves reproducibly at arginines and lysines. The resulting peptides are then ionized to produce charged (protonized) molecules, and travel through a mass analyzer and then to a mass detector. The two ionization techniques most commonly used are (1) matrix-assisted laser-desorption ionization (MALDI) and (2) electrospray ionization (ESI). With MALDI, proteins and peptides are mixed with an energy-absorbing matrix (eg, cinnamic acid) and then are ionized using a laser. With ESI, a sample is introduced in liquid form, and the application of a very high voltage forms a fine spray through a hypodermic needle to ionize peptides. An electromagnetic field causes the ionized peptides to travel through the mass analyzer to a detector. Mass analyzers include time-of-flight (TOF), ion traps, Fourier transform and quadrupoles or combinations (eg, quadrupole-TOF). In TOF MS, the time that it takes the peptide to reach the detector is converted into a mass/charge ratio and is visualized as a mass spectrum. MALDI is generally paired with TOF analyzers (MALDI-TOF). This is excellent for high-throughput studies and is somewhat forgiving of contaminants such as salts, although only single proteins or very simple mixtures can be analyzed. In tandem MS using an ion trap, peptides of a single mass are "trapped" in an electromagnetic field, fragmented, and the resulting fragments are detected and a second mass spectrum is formed. Similarly, fragmentation can occur in a quadrupole prior to TOF during a tandem MS experiment in a quadrupole TOF. Because the fragmentation of a peptide often results in sequence information, a single, high-quality peptide can often be used to identify a protein. In addition, this information can be used to localize posttranslational modifications. However, long run times make these instruments medium-throughput to low-throughput devices, and ESI is generally not as tolerant of salts as MALDI. Therefore, a typical workflow may include 2-DE followed by MALDI-TOF for high-throughput, single-gel-spot analyses. When additional sensitivity, complex sample analysis, or sequence information is required, ESI-MS/MS can be used.


Figure 1
View larger version (40K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1.. An example of a proteomics workflow. A typical proteomics workflow is shown, beginning with the separation of a complex mixture of proteins using 2-DE. Prior to analysis by MS, proteins are excised from the gel (1a), denatured to their primary, linear structure (1b), and digested with a protease such as trypsin (1c). Proteolysis results in predictable protein fragments; for example, trypsin reproducibly cleaves after arginines and lysines. Following protein digestion, peptides are spotted onto a MALDI target plate and introduced into the mass spectrometer via a laser (2a) where they are separated according to their mass. During tandem MS, a peptide may be isolated and undergo fragmentation and a second round of detection, which can result in information on the amino acid sequence of the peptide (2b). Following MS, a mass list is generated and a protein database is queried using a software program called a search engine (3). During database searching, the actual mass list is compared to theoretical digests of all proteins in the database, a process called peptide mass mapping. User-defined parameters, such as limitations on species and mass error rates, help to narrow the search. Following database searching, a list of one or more possible protein candidates is generated, along with a score that indicates the level of significance.

 
Multiple strategies can be used to obtain quantitative data on protein expression (Fig 2 ). In 2-DE and DIGE experiments, proteins are quantitated based on the absorption or fluorescence of the labeling dyes. In mass profiling (eg, SELDI), proteins are quantitated using the total ion current of the peak compared to all of the peaks of the entire spectrum. These quantitative methods are relative (ie, expressed as a ratio to the sum of the signal for all proteins in the sample). In order for MS to be absolutely quantitative, one needs an internal reference standard protein included in each experiment. This is typically done by labeling the proteins with isotope-coded affinity tags or radioisotopic labels such as deuterium and 18O prior to introduction into the mass spectrometer. Although MS-based protein quantification techniques are evolving, they are not as rapid as more traditional approaches, such as enzyme-linked immunosorbent assay, or new technologies, such as photoaptamer arrays8; however, unlike the situation with an MS-based approach, these high-throughput approaches cannot be used to identify and quantitate unknown proteins or posttranslational modifications (eg, proteolysis and glycosylation) and are generally limited by the number of proteins that can be quantitated.


Figure 2
View larger version (37K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2.. Quantitative Proteomics Strategies. Top left, A: postmetabolic labeling strategies label specific amino acid residues following biological experiments. In most cases, proteins are solubilized, and a label, or tag, is introduced. Following digestion with a protease, usually trypsin, tagged peptides are isolated and analyzed by MS. Software is then used to determine the ratio between tagged peptides from different samples. Top right, B: DIGE labels proteins from two or three distinct samples with fluorescent dyes. All samples are subsequently run on a two-dimensional gel and compared using quantitative software. In this example, ELF from a normal person was labeled with one dye (green), and that from a person with acute lung injury was labeled with another dye (red). Green spots represent a relative abundance in the normal person, red spots represent a relative abundance in the patient with acute lung injury, and yellow spots represent no difference. Numbers represent previously identified proteins.16 Bottom left, C: antibody arrays are used to quantitatively analyze several different proteins in a high-throughput manner using small sample volumes. Similar to genomic microarrays, antibody array chips generally have several replicates, and positive and negative controls. Bottom right, D: semi-quantitative software such as Spectrum Mill (Agilent Technologies; Palo Alto, CA), shown here, is increasingly used to screen samples for apparent differences. In this example, an immunodepleted sample (FLOTHR2) is compared to whole plasma (PLASMA) to confirm the removal of albumin. A color-coding system enables the user to determine the presence and identify of a protein, and also the relative abundance of the protein. In this example, the darker the color, the more peptides were detected by the mass spectrometer, and information on the number of spectra and their intensity is also displayed. AA = amino acid coverage.

 
Bioinformatics
Searching Databases To Make Protein Identifications: With improvements in genomic and protein databases (Table 1 ), and with more powerful computational searching techniques (ie, search engines), MS-based approaches are now the method of choice for most protein identification; however, the sensitivity still does not approach that of Western blotting. Utilizing user-defined parameters, the search engine constructs a theoretical digest of proteins in the database and then attempts to match these to peptides from the actual experiment based on their mass. The higher the mass accuracy of the MS, the greater the certainty in protein identifications. For instance, there may be 1,000 peptides in a database with a mean mass of 1,234.5678 ± 50 Da, but only 1 peptide with a mean mass of 1234.5678 ± 0.001 Da. To help the user ascertain the quality of the identifications, a score is assigned to each match. SEQUEST (Thermo Electron Corporation; Waltham, MA) and Mascot (Matrix Science, Ltd; London, UK) are examples of search engine programs that interface experimental MS spectra with the theoretical spectra derived from protein databases. SEQUEST correlates uninterpreted tandem mass spectra of peptides with amino acid sequences from protein and nucleotide databases to determine the amino acid sequence. Mascot integrates peptide mass fingerprinting, sequence query, and tandem MS ion search methods to identify proteins from primary sequence databases.


View this table:
[in this window]
[in a new window]

  		Table 1.. Databases Used for Proteomics Searches

 
Selecting Biomarkers From Proteomics Data (Data Mining): Biomarkers are biological molecules that correlate with disease. The ultimate goal of most clinical proteomics experiments is to identify biomarkers; however, this is not a trivial task since most proteomics experiments typically quantitate hundreds to thousands of proteins, yet few of these proteins are likely to be biomarkers for disease. Furthermore, humans have a large variability in interindividual protein expression. Thus, hundred to thousands of subjects may be needed to identify valid biomarkers, whereas mouse and cell culture experiments typically require only several subjects per group. Since clinical proteomics experiments may result in tens of thousands of data points, sophisticated data exploration algorithms are needed to sift through these data. Commonly used methods for data exploration are hierarchical clustering, K-means, and self-organizing maps.9 Clustering is the organization of a collection of unlabeled patterns into clusters based on similarity, so that patterns within the same cluster are more similar to each other than they are to a pattern belonging to a different cluster. Other techniques that can be used are classification methods such as artificial neural networks, Bayesian networks, classification and regression trees (CARTs), genetic algorithms, statistical pattern recognition, support vector machines, and visualization. Instead of learning functional classifications of proteins in an unsupervised fashion like clustering, classification techniques start with a collection of preclassified samples, with the goal being to learn a classification model that will be able to classify a future sample with corresponding error rates that are as low as possible (Fig 3 ). Classification methods (particularly CARTs) are well-suited for clinical studies because of their straightforward interpretability. A common problem for all of these modeling strategies is overfitting the data (ie, using a large number of variables in the prediction model)10 and lack of validation with an independent data set.


Figure 3
View larger version (24K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3.. CARTs. The CARTs can be used to identify a panel of biomarkers that can be used for the classification of subject. In this example, five biomarkers were selected using a SELDI proteomics plasma study to distinguish a COPD subject from a non-COPD subject. Each node is sequentially labeled and shows splitting criteria based on a SELDI quantification of the protein peak. For instance, M8170 82 ≤ 1.586 would mean that subjects with peak intensities of ≤ 1.586 at m/z ratio at 8170.82 d would move down the left side, and all other subjects would move down the right side. A misclassification is defined when after "dropping" a case down a tree (ie, following the classification rules of a tree), the case is misclassified as COPD when it is a control or as a control when it is COPD. Subjects continue down the tree until they reach terminal nodes. The number of samples that are at each node is given for both the COPD group (black bars) and the control or no-COPD group (gray bars). A terminal node is classified as COPD if the majority of samples in the terminal node are from COPD patients. Otherwise, the terminal node is classified as control (no COPD). Adapted from Bowler et al.24

 
Proteomics in Clinical and Basic Pulmonary Research

The majority of pulmonary proteomic studies have been descriptive and have focused on the following major compartments in the lung: ELF, lung tissue, airway cells, and blood. For more detail, readers are referred to more recent exhaustive reviews.1112

The ELF Proteome
ELF represents the thin layer of fluid covering airway epithelial cells and is readily obtained as BAL fluid (BALF). Newer bronchoscopic microsampling techniques that do not dilute ELF may supplant the BALF procedure for proteomics experiments.13 One of the first studies of the BALF proteome was published in 197914 and used 2-DE to identify BALF proteins in patients with alveolar proteinosis. These early studies were typically limited in their ability to identify specific proteins; however, with recent MS advances, Guo et al15 have used one-dimensional electrophoresis and 2-DE LC coupled to a mass spectrometer to identify 297 unique proteins in the mouse BALF proteome. Most proteins in BALF are plasma proteins such as albumin and Ig; however, a small subset of proteins such as surfactant protein-A and glutathione-S-transferase are significantly more abundant in BALF compared to plasma or serum.1617 Descriptive BALF proteomics studies1112 have been published in healthy subjects, smokers, sarcoidosis, cystic fibrosis, pulmonary fibrosis, asbestosis and mesothelioma, hypersensitivity pneumonitis, immunosuppression, ozone exposure, and acute lung injury.

Proteomics of Lung Cells and Tissue
BAL can also be used to obtain alveolar macrophages for functional proteomics studies. For example, alveolar macrophages have a distinct proteomic profile that relates to their physiologic role in proteolysis, actin reorganization, and cellular adaptation compared to those of blood mononuclear cells.1819 Obtaining lung tissue requires invasive sampling techniques, and the majority of studies have focused on the differences between the normal lung and lung cancers. These studies have relied on both 2-DE and MALDI-TOF approaches. For instance, Yanagisawa et al20 examined 79 lung tumors and tissue from 14 normal lungs using MALDI-TOF and found 15 peaks that could distinguish non-small cell lung cancer patients with a good prognosis from those with a poor prognosis. Three of these peaks were identified as small ubiquitin-related modifier-2 protein, thymosin-ß4, and ubiquitin. MALDI-TOF has also recently been used21 to identify proteomic patterns that differentiate invasive lesions from normal bronchial epithelium. A disadvantage to many of these studies is that they included the surrounding stromal tissue. To overcome this limitation, laser capture microdissection of tumor specimens has been used to improve the specificity of 2-DE for biomarker discovery.22 Although many of these studies have identified promising biomarkers for lung cancer, large validation studies will be required before the biomarkers can be used clinically.

The Blood (Plasma and Serum) Proteome in Pulmonary Diseases
Serum and plasma proteomics are attractive areas of study because they are less invasive than other methods. The SELDI approach has been used to study plasma profiles of lung cancer23 and COPD.24 Other methods, such as 18O metabolic labeling, have recently been used25 to identify 211 proteins that were up-regulated and 246 proteins that were down-regulated in the sera of mice with lung adenocarcinomas (Lewis lung carcinoma), including vascular endothelial growth factor receptor 1.

Emerging Proteomic Technologies

Proteomic technologies are rapidly evolving new ways to look at the proteome. For example, the study of phosphorylated proteins, or phosphoproteomics, not only provides an indication of what proteins and pathways are involved in a particular disease, but can indicate which proteins are likely drug targets.26 More global approaches include phosphoprofiling, which combines enrichment using immobilized metal affinity chromatography with differential labeling of proteins by esterification for quantitative analysis. Various proteins have been detected in exhaled breath condensate by 2-DE,27 and exhaled breath condensate might be a noninvasive source for monitoring respiratory diseases. Finally, several respiratory diseases such as asthma, cystic fibrosis, and bronchitis are characterized by quantitative and qualitative changes in glycoproteins in which samples are enriched for glycoproteins using lectin.28

Conclusion

Although comparative and quantitative proteomics studies may be more technically difficult compared to gene expression studies, they provide orders of magnitude more qualitative information. As with genomics, all proteomics techniques have the advantage of being able to simultaneously study a subset of all proteins as opposed to a single protein. Although clinical proteomics is a technology that is still in its infancy, it has great potential to improve our understanding and treatment of lung disease by identifying patterns of protein expression. These protein expression profiles can reveal broad pathologic processes such as altered proteolytic processing or glycosylation that may not have been evident with other technologies or may reveal complex patterns that can serve as new diagnostic tools (eg, early cancer detection). However, the routine clinical use of proteomic technologies is likely several years away, since, as with other clinical studies, large studies will be needed to validate the clinical utility of each pattern.

Footnotes

Abbreviations: BALF = BAL fluid; CART = classification and regression tree; 2-DE = two-dimensional electrophoresis; DIGE = differential in-gel electrophoresis; ELF = epithelial lining fluid; ESI = electrospray ionization; LC = liquid chromatography; MALDI = matrix-assisted laser-desorption ionization; MS = mass spectrometry; PAGE = polyacrylamide gel electrophoresis; pI = isoelectric point; SELDI = surface-enhanced laser desorption ionization; TOF = time of flight

This work was supported by Flight Attendant Medical Research Institute (R.P.B.) and the Kenneth W. Monfort Program for Research in COPD (M.C.E.).

The authors have reported to the ACCP that no significant conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.

Received for publication February 6, 2006. Accepted for publication May 7, 2006.

References

  1. Lader, EW, Cannon, CP, Ohman, EM, et al (2004) The clinician as investigator: participating in clinical trials in the practice setting; appendix 1: fundamentals of study design. Circulation 109,e302-e304[Free Full Text]
  2. Baggerly, KA, Morris, JS, Coombes, KR Reproducibility of SELDI-TOF protein patterns in serum: comparing datasets from different experiments. Bioinformatics 2004;20,777-785[Abstract/Free Full Text]
  3. Patton, WF Detection technologies in proteome analysis. J Chromatogr B Analyt Technol Biomed Life Sci 2002;771,3-31[CrossRef][Medline]
  4. Jin, M, Drwal, G, Bourgeois, T, et al Distinct proteome features of plasma microparticles. Proteomics 2005;5,1940-1952[CrossRef][ISI][Medline]
  5. Aebersold, R, Mann, M MS-based proteomics. Nature 2003;422,198-207[CrossRef][Medline]
  6. Pusch, W, Flocco, MT, Leung, SM, et al MS-based clinical proteomics. Pharmacogenomics 2003;4,463-476[CrossRef][ISI][Medline]
  7. Aldred, S, Grant, MM, Griffiths, HR The use of proteomics for the assessment of clinical samples in research. Clin Biochem 2004;37,943-952[CrossRef][ISI][Medline]
  8. Smith, D, Collins, BD, Heil, J, et al Sensitivity and specificity of photoaptamer probes. Mol Cell Proteomics 2003;2,11-18[Abstract/Free Full Text]
  9. Orengo, C, Jones, D, Thornton, JM Bioinformatics: genes, proteins and computers. 2003 BIOS Scientific. Oxford, UK:
  10. Hawkins, DM The problem of overfitting. J Chem Inf Comput Sci 2004;44,1-12[CrossRef][ISI][Medline]
  11. Hirsch, J, Hansen, KC, Burlingame, AL, et al Proteomics: current techniques and potential applications to lung disease. Am J Physiol Lung Cell Mol Physiol 2004;287,L1-L23[Abstract/Free Full Text]
  12. Wattiez, R, Falmagne, P Proteomics of bronchoalveolar lavage fluid. J Chromatogr B Analyt Technol Biomed Life Sci 2005;815,169-178[ISI][Medline]
  13. Ishizaka, A, Watanabe, M, Yamashita, T, et al New bronchoscopic microsample probe to measure the biochemical constituents in epithelial lining fluid of patients with acute respiratory distress syndrome. Crit Care Med 2001;29,896-898[CrossRef][ISI][Medline]
  14. Bell, DY, Hook, GE Pulmonary alveolar proteinosis: analysis of airway and alveolar proteins. Am Rev Respir Dis 1979;119,979-990[ISI][Medline]
  15. Guo, Y, Ma, SF, Grigoryev, D, et al 1-DE MS and 2-D LC-MS analysis of the mouse bronchoalveolar lavage proteome. Proteomics 2005;5,4608-4624[CrossRef][ISI][Medline]
  16. Bowler, RP, Duda, B, Chan, ED, et al Proteomic analysis of pulmonary edema fluid and plasma in patients with acute lung injury. Am J Physiol Lung Cell Mol Physiol 2004;286,L1095-L1104[Abstract/Free Full Text]
  17. Sabounchi-Schutt, F, Astrom, J, Eklund, A, et al 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[CrossRef][ISI][Medline]
  18. Jin, M, Opalek, JM, Marsh, CB, et al Proteome comparison of alveolar macrophages with monocytes reveals distinct protein characteristics. Am J Respir Cell Mol Biol 2004;31,322-329[Abstract/Free Full Text]
  19. Wu, HM, Jin, M, Marsh, CB Toward functional proteomics of alveolar macrophages. Am J Physiol Lung Cell Mol Physiol 2005;288,L585-L595[Abstract/Free Full Text]
  20. Yanagisawa, K, Shyr, Y, Xu, BJ, et al Proteomic patterns of tumour subsets in non-small-cell lung cancer. Lancet 2003;362,433-439[CrossRef][ISI][Medline]
  21. Rahman, SM, Shyr, Y, Yildiz, PB, et al Proteomic patterns of preinvasive bronchial lesions. Am J Respir Crit Care Med 2005;
  22. Ahram, M, Best, CJ, Flaig, MJ, et al Proteomic analysis of human prostate cancer. Mol Carcinog 2002;33,9-15[CrossRef][ISI][Medline]
  23. Xiao, X, Liu, D, Tang, Y, et al Development of proteomic patterns for detecting lung cancer. Dis Markers 2003;19,33-39[ISI][Medline]
  24. Bowler, RP, Canham, ME, Ellison, MC Surface enhanced laser desorption/ionization (SELDI) time-of-flight MS to identify patients with chronic obstructive pulmonary disease. COPD: J Chron Obst Pulm Dis 2006;3,41-50[CrossRef]
  25. Hood, BL, Lucas, DA, Kim, G, et al Quantitative analysis of the low molecular weight serum proteome using 18O stable isotope labeling in a lung tumor xenograft mouse model. J Am Soc Mass Spectrom 2005;16,1221-1230[CrossRef][ISI][Medline]
  26. Lim, YP Mining the tumor phosphoproteome for cancer markers. Clin Cancer Res 2005;11,3163-3169[Abstract/Free Full Text]
  27. Aulak, KS, Miyagi, M, Yan, L, et al Proteomic method identifies proteins nitrated in vivo during inflammatory challenge. Proc Natl Acad Sci U S A 2001;98,12056-12061[Abstract/Free Full Text]
  28. Qiu, R, Regnier, FE Comparative glycoproteomics of N-linked complex-type glycoforms containing sialic acid in human serum. Anal Chem 2005;77,7225-7231[Medline]



This article has been cited by other articles:


Home page
 ChestHome page
E. Kipnis, K. Hansen, T. Sawa, K. Moriyama, A. Zurawel, A. Ishizaka, and J. Wiener-Kronish
Proteomic Analysis of Undiluted Lung Epithelial Lining Fluid
Chest, August 1, 2008; 134(2): 338 - 345.
[Abstract] [Full Text] [PDF]

Home page
 Eur Respir JHome page
R. Farre and A. T. Dinh-Xuan
Translational research in respiratory medicine
Eur. Respir. J., December 1, 2007; 30(6): 1041 - 1042.
[Full Text] [PDF]

Home page
 ChestHome page
E. Kipnis, B. Guery, and J. Wiener-Kronish
Promises and Limitations of the Bronchoscopic Microsampling Probe
Chest, October 1, 2007; 132(4): 1414 - 1414.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (7)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bowler, R. P.
Right arrow Articles by Reisdorph, N.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bowler, R. P.
Right arrow Articles by Reisdorph, N.
Related Content
Right arrow Translating Basic Research into Clinical Practice

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS