Toward functional proteomics of alveolar macrophages

Haifeng M. Wu, Ming Jin, Clay B. Marsh

Abstract

Alveolar macrophages (AM) belong to a phenotype of macrophages with distinct biological functions and important pathophysiological roles in lung health and disease. The molecular details determining AM differentiation from blood monocytes and AM roles in lung homeostasis are largely unknown. With the use of different technological platforms, advances in the field of proteomics have made it possible to search for differences in protein expression between AM and their precursor monocytes. Proteome features of each cell type provide new clues into understanding mononuclear phagocyte biology. In-depth analyses using subproteomics and subcellular proteomics offer additional information by providing greater protein resolution and detection sensitivity. With the use of proteomic techniques, large-scale mapping of phosphorylation differences between the cell types have become possible. Furthermore, two-dimensional gel proteomics can detect germline protein variants and evaluate the impact of protein polymorphisms on an individual's susceptibility to disease. Finally, surface-enhanced laser desorption and ionization (SELDI) time-of-flight mass spectrometry offers an alternative method to recognizing differences in protein patterns between AM and monocytes or between AM under different pathological conditions. This review details the current status of this field and outlines future directions in functional proteomic analyses of AM and monocytes. Furthermore, this review presents viewpoints of integrating proteomics with translational topics in lung diseases to define the mechanisms of disease and to uncover new diagnostic and therapeutic targets.

  • monocytes
  • chronic obstructive pulmonary disease
  • phosphoproteomics
  • protein polymorphism
  • subproteomics

BACKGROUND

the mononuclear phagocyte system consists of bone marrow monoblasts, promonocytes, blood monocytes, and tissue macrophages. Under normal circumstances, monocytes derived from bone marrow circulate in the blood for <48 h and enter tissue compartments to differentiate into macrophages. Compared with monocytes, macrophages have a longer life span and heterogeneous functions (19, 22). Macrophages in different anatomical locations express different biological activities. These findings suggest that the tissue's microenvironment helps control macrophage differentiation and the biological functions unique to the tissue compartment. However, the cellular mechanisms determining these processes are largely unknown.

Alveolar macrophages (AM) are one unique class of macrophage that function primarily in lung defense against inhaled particulate matter, microorganisms, and environmental toxins (19, 22). These cells possess Fcγ and complement receptors that respond to the Fc portion of IgG and C3 components, respectively. These receptors enable AM to participate in initial host defenses through the production of chemokines such as IL-8 (44) and monocyte chemoattractant protein-1 (45) and growth factors, such as macrophage colony-stimulating factor, which augments monocyte survival (37, 70). In addition, these cells have many other diverse roles in the regulation of inflammation and adaptive immune response, including antigen presentation and the production of reactive oxygen and nitrogen species and metalloproteases (3, 59, 64). Understanding differences in protein expression between AM and their precursor blood monocytes may give insight into biological differences in these cells (35).

AM play a critical role in the pathogenesis of numerous pulmonary diseases. For instance, AM accumulate in the alveoli in smokers and display a direct correlation to the severity of chronic obstructive pulmonary disease (COPD) (21, 34). The activation of AM by smoking releases proteases, inflammatory mediators, and reactive oxygen species, all of which provide models of cellular mechanisms leading to injuries of the lung parenchyma in COPD (35, 10, 40, 42, 51, 56, 57, 60, 67). AM are also implicated in the pathogenesis of idiopathic pulmonary fibrosis, asthma, and pulmonary sarcoidosis (16, 17, 48, 53, 73, 75, 80). Therefore, a proteomic analysis of AM in respective patient populations may provide new insight into the mechanisms underlying inflammatory lung diseases.

The term “proteome” refers to all measurable proteins and their levels in cells under a given condition. “Proteomics” applies the integration of changes in the proteome to different pathophysiological states. Thus proteomics offers the opportunity to understand disease processes and to develop new biomarkers for early diagnosis/disease prediction (26, 29). Yet, proteomics involves various platforms of technologies that are technically challenging, each with advantages and disadvantages. In this manner, a focused proteomic research effort on a particular cell type or disease may offer substantial advances in the field. This review introduces technologies and methods available in the proteomic analysis of AM and monocytes and describes the data-driven strategy to generate specific hypotheses (35). Furthermore, this review will demonstrate the powerful application of proteomics in discovering germline protein polymorphisms in mononuclear phagocytes. This new finding opens a new paradigm to study the impact of protein polymorphisms on individual susceptibility to lung diseases.

BASIC CONSIDERATIONS OF AM PROTEOMIC RESEARCH

In comparison to genomics, the field of proteomics is particularly challenging because of the wide dynamic range of protein abundance, complicated protein properties (such as mol wt, isoelectric point, extent of hydrophobicity, and posttranslational modifications), and the complexity of protein-protein interactions. Several proteomic technologies have come to light, each with their unique strengths and limitations: 1) two-dimensional (2D) gel electrophoresis, coupled with mass spectrometry for the determination of protein identity, is currently the workhorse of proteomics due to decades of experience with 2D gels and the recent development of an immobilized pH gradient for isoelectric focusing electrophoresis, and 2) surface-enhanced laser desorption and ionization (SELDI) provides rapid protein profiling analysis that is particularly suitable for the study and validation of disease markers in large numbers of clinical specimens. Other proteomic tools such as high-throughput mass spectrometry, protein or antibody arrays, and multidimensional chromatography using microfluidity technology, are also emerging and will likely provide powerful alternatives in the future (20, 24, 72, 79, 78). Current opinion holds that no single technology will meet the needs of all types of proteomics-based investigations. Yet, the implementation of multiple strategies (e.g., subproteomics and phosphoproteomics) and various technology platforms will likely provide more comprehensive proteomic analyses in complex tasks like disease proteomics (26).

Protein resolution is one of the key issues during 2D gel analysis. In general, the larger the gel format, the better the proteins are resolved. Consistency in gel format and in the amount of protein loaded onto each 2D gel are recommended for clinical samples or samples involving the same cell type. This consistency reduces variability and enables the gel images to be used prospectively for long-term research. As a result, gel images can be compared with each other to evaluate biological variations of proteomes, potential protein polymorphisms between individuals, or an individual's heterogeneous response to drug treatment. The largest commercial 2D gel format (20 × 26 cm) allows for loading a maximum of 2 mg of proteins. The resolving power for this type of gel is ∼1,300–1,800 AM protein spots on each 2D gel under pH 3–10 nonlinear on the first dimension and 12% SDS-PAGE on the second dimension. Again, a fixed amount of protein is required for loading onto each 2D gel to ensure accurate comparison of 2D gels between samples. Reproducibility is another key issue in 2D gel proteomics. It is particularly challenging because 2D gel analyses in their entirety involve numerous technically difficult steps. Variability between experiments introduces artifacts to the final result. To augment reproducibility, it is necessary to standardize the procedures of proteomic analysis such as cell sample preparation, conditions for first and second dimensions of electrophoresis, gel fixing, gel staining, and image acquisition. Good reproducibility is accepted to have an intra-assay coefficient of variation of <30% in protein quantification for multiple runs using the same sample (49). Furthermore, success in reproducibly self-casting second-dimensional SDS polyacrylamide gels reduces the cost of each 2D experiment and provides flexibility in altering polyacrylamide gel percentages for second dimension. Such alterations increase the resolution power by focusing only on subsets of a total proteome during subproteomic analysis as described later.

High protein resolution and high reproducibility of 2D gels are keys to successful proteomic analyses. Figure 1 illustrates the importance of these factors in such analyses. In this study, U-937 monocytic cells were treated with eight different conditions followed by 2D gel analysis. Protein spots in each 2D gel were matched to corresponding spots in the U-937 cell reference 2D map and quantified (each spot quantified as a percentage of total protein intensity in a given 2D gel). Excellent reproducibility in protein patterns and protein density between all eight gels is displayed in the close-up view of a gel area from all eight 2D gels in this experiment. This area also contains a candidate spot that demonstrates significant quantity changes in various treatment conditions (quantification of this candidate spot is shown in Fig. 1C). Mass spectrometry analysis identified this spot as IL-1β. Thus highly reproducible results combined with the resolution power of 2D gels enables an accurate examination of cellular proteome changes in various pathophysiological conditions. This particular experiment demonstrated that the 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitor simvastatin (statin) inhibited LPS-mediated IL-1β production in this monocytic cell line, thus providing new insight into the anti-inflammatory property of statin.

Fig. 1.

Resolution power and reproducibility of two-dimensional (2D) gel proteomics. U-937 monocytic cells were treated with 5 mM simvastatin (marked as statin), 1 mM p38 MAPK inhibitor (SB-203580, marked as SB80), or both at 37°C for 1 h before LPS stimulation for 12 h. The cells (3 × 107 cells) were lysed in 2D lysis buffer and subjected to 2D gel electrophoresis using pH 3–10 immobilized pH gradient strips for first-dimension and 12% SDS-PAGE for second-dimension electrophoresis. A: 1 of 8 2D gels (20 × 26 cm) in this experiment. The box represents the gel area to be illustrated. B: close-up view of a particular gel area from 8 2D gels in this experiment. Arrows indicate 1 candidate spot that exhibits quantitative changes in 8 different treatment conditions. Note that the remaining protein spots remain largely unchanged in all 8 experimental conditions. C: quantification of this candidate spot in the 8 2D gels. Means ± SD were derived from 3 separate experiments. Mr, molecular weight.

To benefit long-term proteomic research, it is necessary to create a standard 2D reference gel for a particular cell type or set of experiments run under the same electrophoretic condition. Usually, the gel with the best resolution among the series is selected as the reference gel. All other 2D gels are matched to this reference 2D gel for spot identification. With good reproducibility between runs, one would expect to have >80% of all protein spots on each 2D gel successfully matched to their respective protein spots on the reference gel. All de novo protein spots on nonreference gels are added to the reference gel image during the matching process. This allows the reference gel to encompass all the protein spots in this set of experiments. Within this set of experiments, all protein identities (ID) obtained by mass spectrometry can be assigned to their respective spots on the 2D reference gel. Thus this reference 2D gel provides available information on protein ID for all subsequent 2D gels matched to the reference 2D gel. As a result, repeated mass spectrometry on the same protein is not required, making the process of proteomic analysis for any macrophage- or monocyte-related proteomic research faster, easier, and less expensive.

It is worth noting that the cell lysis buffer contains high concentrations of urea, thiourea, and reducing agents that interfere with most assays for measuring protein concentration. Therefore, a small aliquot of cells is recommended to be lysed in an appropriate lysis buffer for protein determination while the majority is dissolved directly into 2D gel cell lysis buffer for proteomic analysis.

SUBPROTEOMICS TO ENHANCE RESOLUTION AND DETECTION OF PROTEINS

Sequencing of the human genome has revealed that there are ∼35,000 genes per cell (39). Because each gene can be differently formed, spliced, or modified to generate multiple protein products, there may be as many as 1,000,000 differently modified proteins in humans (39). Most of these proteins and their functions remain unknown at the present time. Subproteomics refers to the analysis of a subset of a total proteome according to either spatial locations or biochemical properties. Available techniques include proteomic analysis of particular subcellular fractions such as cell membrane or proteomic analysis of a fraction of proteins separated based on isoelectric point (pI), molecular weight (Mr), or binding properties. In this way, interesting target protein(s) are greatly enriched or better resolved, either before or during 2D gel analysis, to provide improved protein detection sensitivity. Two routine methods are illustrated that demonstrate the plausibility of subproteomic analysis to enhance proteomic analysis.

One example is subproteomics using a narrower pH range for the first electrophoretic dimension as illustrated in Fig. 2. Here, 2D gels of U-937 monocytic cell lysates were performed at two different pH ranges on first-dimension electrophoresis (pH 3–10 vs. pH 5–6). As demonstrated, the first dimension conducted under pH 5–6 provided better separation of all proteins with a pI between 5 and 6. As a result, subproteomics serve as a feasible tool to simultaneously increase resolution power and detection sensitivity by focusing only on subsets of a total proteome. Similarly, one can change the percentage of polyacrylamide in second-dimension electrophoresis to hone in on a particular fraction of proteins with a defined range of molecular mass.

Fig. 2.

Subproteomics using different pH ranges on first dimension. A: 2D electrophoresis of cellular proteins derived from U-937 monocytic cells on a broader pH range of 3–10 during first-dimension electrophoresis. B: 2D electrophoresis of cellular proteins on a narrower pH range of 5–6 used on first dimension. All the proteins in B represent the proteins in the box in A (arrow). Solid vertical protein aggregates at both sides of pH 5–6 gel represent those proteins with pI >6 or pI <5.

Figure 3 illustrates the results of subcellular proteomics where proteomic analysis is conducted after subcellular fractions are obtained by differential centrifugation techniques. Both whole cell lysates and organelle fractions were prepared from U-937 monocytic cells. A total of 2 mg of cellular proteins were loaded onto each 2D gel (pH 3–10 on first dimension and 10% on the second dimension). When organelle proteins are compared with whole cell lysates, there are striking differences in protein patterns and density between these two samples. Many proteins not apparent in whole cell lysates constituted the major protein components in an organelle proteome. This demonstrated the power of subcellular proteomics to analyze a subset of proteins uncovered during global 2D gel analysis of whole cell lysates.

Fig. 3.

Illustration of subcellular proteomics. U-937 monocytic cells were used in this experiment to prepare for whole cell lysates and organelle proteins. 2D gels of organelles and whole cell lysates were compared. 2D gel of whole cell lysates is shown at left and that of organelles is at right. The enclosed boxes represent a portion of the gel as magnified. Many proteins not visible in the 2D gel of whole cell lysates become apparent in the gel of organelles (gray arrows). Black arrows indicate proteins that appear in both types of 2D gels.

PHOSPHOPROTEOMICS

Posttranslational protein modification by phosphorylation plays a central role in transducing signals responsible for a variety of biological processes. Approximately one-third of all proteins in eukaryotic cells are phosphorylated at any given time (15, 43). Protein phosphorylation can occur at tyrosine, serine, or threonine residues, each regulating respective cellular events. In macrophages, the expression of cellular activities involves numerous protein kinases and their downstream-phosphorylated substrates (30, 50, 55). Complete phosphoproteomics of a cell involves identification of all phosphoproteins, quantification of each phosphorylation event, and localization of the exact residues that are phosphorylated. The first two steps in this analysis focus on screening and discovering potential protein targets related to different pathophysiological conditions. The third step, defining the site of a protein's phosphorylation reaction, however, requires a detailed analysis of the individual target spot using matrix-assisted laser desorption/ionization time-of-flight and/or tandem mass spectrometry. This last step will not be discussed in this review. Readers should refer all questions concerning this step to recent publications listed in the references (2, 23, 38).

Two different methods can be used to analyze phosphorylation proteomes and to identify target proteins of interests in mononuclear phagocytes. One method is phosphoprotein staining that uses Pro Q fluorescence dye to detect the general phosphorylation proteome. This stain allows detection of phosphorylated amino acids with a detection sensitivity of 0.1 fmol of monophosphorylated proteins (62, 66). The procedure is easy to perform and offers a feasible approach to detecting general protein phosphorylation states at either tyrosine, serine, or threonine residue(s).

As shown in Fig. 4, both monocytes and AM reveal ∼600 phosphorylated proteins each but show distinct and different patterns in the proteins that are phosphorylated. Similarly, protein patterns stained by Sypro Ruby also revealed differences between the two cell types. One advantage of Pro Q staining is that the same 2D gel can be initially stained with Pro Q to measure the protein phosphorylation proteome and then later stained with Sypro Ruby to measure total protein levels. Another unique feature is that many phosphorylated proteins, positively identified by Pro Q stain, are not apparent with Sypro Ruby stain, indicating detection sensitivity of Pro Q diamond exceeds that of Sypro Ruby fluorescence dyes. Whereas the Pro Q staining technology is very sensitive in assessing protein phosphorylation states, the protein ID process has sometimes proven to be difficult due to the detection of phosphorylated proteins at very low protein abundance.

Fig. 4.

Illustration of phosphorylation proteome differences between alveolar macrophages (AM) and blood monocytes. AM and monocyte lysates obtained from 1 donor were analyzed by 2D gels initially stained with Pro Q diamond dye to obtain a phosphorylation proteome and then stained with Sypro Ruby to examine total proteome. Comparison of both the phosphorylation proteome and the total protein proteome between monocytes and AM were illustrated by showing the same gel area. Small black arrows point to proteins that are visible in both proteome and phosphoproteome of all 4 gels, indicating that the same gel areas are compared among 4 gel images. Larger arrows with boxes show magnified areas.

The second approach in evaluating the phosphoproteome is group-specific phosphoproteomics. This approach combines 2D gel electrophoresis with Western blot analysis to examine the tyrosine or serine phosphorylation proteomes. The goal here is to determine specific protein residue(s) that undergo phosphorylation modification. Figure 5 illustrates one example using a monoclonal anti-phosphotyrosine antibody. With the use of image analysis software, the immunoreactive protein spots from control and LPS-stimulated cells were matched, the spots were quantified, and the differences were analyzed. Similarly, 2D gel analysis can be combined with Western blot using anti-phosphoserine antibody (6) to measure changes in serine phosphorylation states (data not shown). These techniques will define the tyrosine phosphorylation proteome and serine phosphorylation proteome in AM and blood monocytes. Once positive immunoreactive spots are identified, the Western blot membrane will then be matched to 2D gels to locate the respective protein spot and excised for protein identity analysis.

Fig. 5.

LPS induced tyrosine phosphorylation changes in blood monocytes. A shows monocytes without LPs treatment. B shows monocytes treated with LPS (100 ng/ml) for 10 min. Whole cell lysates were then analyzed by 2D gel, followed by Western blot using anti-phosphotyrosine monoclonal antibody. Tyrosine phosphorylation proteomes were compared between 2 treatment groups. Two examples are illustrated in each type of phosphorylation change: black arrows in both A and B represent proteins with no change in tyrosine phosphorylation. Gray arrows in B represent an LPS-induced increase in tyrosine phosphorylation. Open arrows in A represent the proteins that show a decrease in tyrosine phosphorylation after LPS treatment.

SELDI

In the SELDI technology, small amounts of biological samples are applied to a ProteinChip with well-defined surface chemistries. Several ProteinChips exist with surfaces that are cationic, anionic, hydrophobic, or ligand (antibody, receptor, DNA, etc.) specific that allow target proteins to be fractionated or enriched in combination with a selected wash condition. This is then followed with on-chip mass profiling and quantification of the proteins bound to each ProteinChip.

One advantage to SELDI is the integration of on-chip capture, purification, and quantitative detection of targeted proteins on a single platform. Additionally, SELDI requires small sample sizes, provides high-resolution power for low-mass protein components, and offers fast protein profiling analysis (14, 32, 77). A close comparison between 2D gel and SELDI, the two most often used proteome analysis systems, is shown in Table 1.

View this table:
Table 1.

Comparison of two frequently used methodologies in proteomic analysis

Meanwhile, Fig. 6 demonstrates the application of SELDI on AM research. Approximately 25–30 protein components with Mr between 7.5 and 17.5 kDa are compared in duplicate between AM and monocytes from the same individual. Examination of cell lysates by all available ProteinChips in different washing conditions will allow ∼300–400 protein peaks to be compared between samples. Although SELDI provides a fast proteome analysis in a semiautomated fashion, the general protein resolution power is not comparable to that of a 2D gel that resolves up to 2,000 protein spots. Another limitation of SELDI includes the inability to directly obtain a protein's ID upon discovery of a target protein peak. To address this issue, a scaled-up protocol exists to purify target proteins based on their binding properties to various chips. The idea is to obtain a protein ID from the purified protein. This scaled protocol is, however, very time consuming. Alternatively, technology is evolving that couples the SELDI technology with tandem mass spectrometers. Interfacing these two instruments will enable on-chip amino acid sequencing and subsequent protein identification (8).

Fig. 6.

Protein profiling analysis of AM and monocyte cell lysates by surface-enhanced laser desorption and ionization (SELDI). Here, cell lysates (∼1 μg) from AM and monocytes of the same donor were compared in duplicate on a weak cationic exchanger (WCX) ProteinChip using SELDI. Distinct protein peaks for each cell type are exemplified by arrows.

INSIGHT OF PROTEOMIC DISCOVERIES INTO AM BIOLOGY

Proteomics allows for large-scale protein analysis and will likely provide new knowledge in phagocyte biology. Table 2 summarizes some protein characteristics that distinguish AM from blood monocytes (35). This information provides new clues into the biology underlying monocyte differentiation into AM. As predicted, many of the proteins expressed by AM correlated to their physiological properties. First, there is generally an increased level of proteinases such as thiol proteases (cathepsins B, D, H, and X) and serine protease tripeptide peptidase I in AM. These observations are consistent with literature reports (11, 47, 58, 61) and additionally provide more information on the number and characteristics of isoforms for each protease. Second, AM and blood monocytes show different expression patterns of actin regulatory elements such as heat shock protein 27, macrophage capping protein, also named Cap G and cofilin. Because actin regulatory elements are critical for phagocytosis, chemotaxis, and cytokine release of phagocytes (1, 9, 25, 36, 46, 54, 71, 74, 76), different expression patterns of actin regulatory proteins between these two cell types may constitute the basis for their different biological activities. Third, there were many unique protein characteristics in AM that were consistent with their cellular adaptation to the alveolar milieu of higher levels of oxygen pressure and reactive oxygen species. For example, superoxide dismutase and peroxiredoxin 6 are significantly elevated in AM. In contrast, blood monocytes displayed an abundance of proteins in transcription, inflammation, and control of proteolysis. Finally, high resolution of these 2D gels revealed many isoforms of the same proteins that may include germline protein variants (GPV), various splicing variants, and posttranslational modifications such as phosphorylation, glycosylation, and proteolytic cleavage. We are currently evaluating the meaning of these modifications to explore the functional implications of these protein isoforms/variants and establish whether each variant is related to disease. Overall, proteomic analyses offer insight on protein characteristics that are potentially fundamental to biology and disease processes. Importantly, the information provided at this global level allows for a rational selection of target proteins for further corroborative studies using specific assays or model systems.

View this table:
Table 2.

Proteins exhibiting significant level changes between AM and monocytes

PROTEOMICS OF AM IN LUNG DISEASES

Proteomics offers tremendous opportunities to understand disease mechanisms and to discover new markers for diagnosis and disease prediction. An example can be seen in COPD. Because AM are strongly implicated in the disease development of COPD (7, 27, 28, 31, 52, 63, 65, 67, 68), proteomic analysis of AM may be a powerful tool to study the role of these cells in the genesis of COPD in people. Situated in the alveoli, AM are exposed to environmental challenge by smoking. As a result, smoking significantly alters many biological functions in AM (7, 27, 28, 31, 52, 63, 65, 67, 68). A slow progression to COPD in smokers parallels the chronic increase of macrophages in alveolar space, suggesting the possibility for a causal relationship between smoking and altered macrophage proteomic expression in humans who develop COPD.

Whereas smoking-induced COPD is a major cause of global morbidity and mortality, the detailed molecular events underlying the pathogenesis of COPD remain elusive. One primary hypothesis suggests an imbalance in protease-antiprotease in the lungs of smokers who develop COPD (35, 10, 12, 13, 33, 40, 41, 51, 56, 60, 67, 69). Other models indicate that smoking induces reactive oxygen species in the lung that, in turn, lead to cellular activation and subsequent tissue damage (42, 57) or that a deficiency in anti-inflammatory cytokines like transforming growth factor-β may lead to lung destruction (51). A systemic proteomic study of AM in smokers who develop COPD vs. smokers who do not will offer new insight on where and how these various pathways converge, resulting in the clinical phenotype of COPD.

To illustrate how proteomics is applied to disease pathophysiology, AM proteome changes by smoking is presented since smoking is a necessary environmental factor for >95% of COPD cases. As shown in Fig. 7, AM proteome from three smokers (>10 years with 1 pack cigarettes/day) were compared with the AM proteome from those of nonsmokers. As predicted, smoking produced many changes in overall protein patterns and quantification in AM 2D gels. Reproducible changes of few specific proteins are given. It should be noted that because of the presence of proteome heterogeneities between individuals, it is critical to define disease-related changes reproducibly present between all subjects. The results demonstrate the feasibility of a proteomic approach to provide more information about the pathogenesis of smoking-related COPD and can be extended to other lung disorders.

Fig. 7.

Smoking-induced specific proteome changes in AM (n = 3). Shown is 1 of 3 2D gels from either smokers or nonsmokers, with a zoomed-in illustration of 1 gel area from all 6 subjects. Boxed areas illustrate the reproducible changes of the AM proteome in smokers (n = 3) compared with the proteomes of nonsmokers (n = 3). Lightly shaded arrows point to protein spots present only in smokers. These proteins are 1 of the actin derivatives (left) and 1 isoform of cathepsin D (right). Heavily shaded arrows illustrate proteins present in both groups. They are galectin-1 (left) and thioredoxin (right).

SELDI can be used as an alternative approach to determine proteome changes in smokers' AM. This is illustrated in Fig. 8 where AM cell lysates from a smoker and nonsmoker were compared using the weak cationic exchanger ProteinChip. Proteins bound to this chip were displayed in Fig. 8. Differences between these two subjects are clearly seen and are exemplified by arrows. These preliminary results demonstrate the feasibility of using SEDLI as a complementary approach to provide additional analysis into smoking-induced proteome changes in AM.

Fig. 8.

SEDLI analysis of smokers' AM lysates. One microgram of AM lysate from a smoker and a nonsmoker were compared by SELDI using a WCX chip. Proteins uniquely present in a smoker (heavily shaded) and a nonsmoker (lightly shaded) are depicted by arrows.

PROTEIN POLYMORPHISMS AND THEIR POTENTIAL IMPACT ON CLINICAL MANAGEMENT OF LUNG DISEASES

Proteomics can influence the clinical practice of lung diseases (29). It provides a powerful tool to discover new disease markers and to study disease pathogenesis. A number of specimens can be collected and analyzed by the proteomic approach. These include bronchoalveolar lavage (BAL) fluid, BAL cells, blood/plasma, or lung biopsy specimen. Additionally, these specimens can be subjected to analyses using different platforms of proteomic technologies or subproteomic analysis.

As AM and monocyte 2D gel databases were constructed in our laboratories, we recognized that many proteins displayed differences in electrophoretic mobility among individuals. Control experiments demonstrated such interindividual differences of proteins were not affected by in vitro stimulation with LPS and other inflammatory cytokines. Staining by specific fluorescence dye of Pro Q for phosphorylated proteins further demonstrated that the protein's electrophoretic mobility difference among individuals is not the result of protein phosphorylation. In contrast, in any given individual, a polymorphic protein's electrophoretic mobility remains identical in both AM and monocyte 2D gels, further demonstrating that the observed protein polymorphism is a genetically determined phenomenon. Such proteins showing constitutionally different electrophoretic properties among individuals are referred as “germline polymorphic proteins.” Figure 9 illustrates macrophage capping protein (Cap G) as one of the examples of a germline polymorphic protein.

Fig. 9.

A: 2D gel of AM. B: gel area from both monocytes and AM from 3 individuals having different germline protein variants (GPV) for macrophage capping protein (Cap G). Lightly shaded arrows point to Cap G proteins showing polymorphic features between 3 subjects. There are 4 isospots for Cap G, constituting 3 patterns of spot distribution, also referred to as GPV. Heavily shaded arrows point to the spots present in both AM and monocyte 2D gels of all 3 individuals indicating that the same gel areas were compared between both cell types and among all individuals. Note that electrophoretic mobility of Cap G isoforms are perfectly aligned between AM and monocytes within the same individual. C: shows a phosphoprotein stain by Pro Q dye. All 4 isoforms of Cap G in subject 1 (left) are unphosphorylated, confirming changes in electrophoretic mobility were not due to phosphorylation events. Asterisk points to proteins with positive stains for protein phosphorylation.

In this review, we propose the term “germline protein polymorphisms” to define a subset of interindividual proteome heterogeneity that was genetically determined. We further propose the term germline protein variant (GPV) to define various variant forms of a protein showing germline polymorphic features. In other words, each GPV refers to 2D isospot patterns of the protein that displays variations on a population scale, but never displays variations within an individual (e.g., AM vs. monocytes). To describe further, each polymorphic protein encompasses several GPV displayed as discrete isospot(s) on a 2D gel but is not a result of in vitro stimulation of the cells. An isospot is one of a set of multiple protein spots with the same protein ID.

It should be noted here that there are other biological variations of proteins among individuals such as protein quantity and type, number of isoforms, and differences in posttranslational modifications. Such variations are likely to be more of a result from environmental stimulation or host-environmental interactions. Germline protein polymorphisms strictly refers to a small portion of proteome heterogeneities among individuals that are genetically determined. However, the exact percentage of cellular proteins demonstrating germline polymorphic features is not known. Initial protein identity determination of 100 AM proteins revealed 3 proteins having these characteristics. This suggests that protein polymorphisms may occur in <5% of cellular proteins. If this is accurate, then GPV typing of polymorphic proteins in each subject can be performed analogous to human leukocyte antigen typing or blood group antigen typing performed in clinical laboratories, as exemplified in Table 3. In this way, the frequency of each GPV for each polymorphic protein can be compared between disease and control groups to establish possible clinical associations.

View this table:
Table 3.

Illustrating GPV typing of polymorphic proteins in mononuclear phagocyte system

An example of a protein showing GPV is Cap G. The frequency of each GPV of Cap G is calculated from the 2D gel database, which currently consists of monocyte 2D gels from 60 subjects (Table 4). Nineteen of these 60 subjects have 2D gels of both monocytes and AM, with the remaining subjects having just 2D gels of either monocytes or AM. Importantly, Cap G is a biologically relevant candidate that is implicated in the regulation of phagocytosis (18, 76). Polymorphic features of Cap G, therefore, may bear a significant biological impact.

View this table:
Table 4.

Frequency of each GPV of the polymorphic protein Cap G

Usually, protein polymorphism is noted when two or more spots on a 2D gel have matching protein ID. Once a candidate polymorphic protein is noted, one can check the candidate spot(s) against the 2D gel image database to determine whether this protein is polymorphic among individuals and to determine the frequency of each GPV. In this light, a focused proteomic research program that has built up a protein ID database and a large 2D gel database for a particular cell type is essential for discovering major germline protein polymorphisms and for investigating their respective impacts in the development of diseases.

The concept of association analysis of protein polymorphism with disease is just evolving due to technological advancements in proteomics. There are four apparent advantages to analyzing protein polymorphism compared with the conventional candidate gene association study: 1) changes in a protein's properties such as pI or Mr will likely bear a direct impact on altering the protein's biological activity; 2) 2D gel proteomic analysis simultaneously examines and establishes protein phenotyping (GPV typing) for all observable germline protein polymorphisms in a single 2D gel experiment; 3) the same 2D gel image can be stored to analyze newly discovered germline protein polymorphisms in the future; and 4) it also offers quantitative measurement of each polymorphic protein on a 2D gel. Thus proteomics provides an unprecedented opportunity to define protein polymorphisms within an individual. Clinical association analyses of each protein polymorphism and each respective GPV with disease will likely identify novel cellular elements responsible for an individual's disease susceptibility and responses to drug therapy.

GRANTS

This work is supported in part by National Heart, Lung, and Blood Institute (NHLBI) Grant K08-HL-03279 (to H. M. Wu), NHLBI Grants HL-63800, RO1-HL-66108, RO1-HL-67176, and PO1-HL-70294 (to C. B. Marsh), and by grant support from Ohio Biomedical Research and Technology Transfer Commission (to H. M. Wu).

REFERENCES

View Abstract