| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
Departments of *Anesthesiology and Critical Care, 
Biochemistry and Biophysics, and the Johnson Research Foundation, University of Pennsylvania, Philadelphia
Address correspondence and reprint requests to Jonas S. Johansson, 319C, John Morgan Building, University of Pennsylvania, 3620 Hamilton Walk, Philadelphia, PA 19104. Address e-mail to JohanssJ{at}uphs.upenn.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionConclusionsReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionConclusionsReferences |
|---|
An understanding of proteins, even more so than genes, holds tremendous clinical promise across every discipline of medicine and certainly in anesthesiology, where most of the major anesthetics interact with and modulate the function of a large number of protein targets with very limited specificity. Anesthesiologists deliver drugs that impact total body function to a degree that remains mechanistically ill-defined but that certainly affects the balance of protein function and expression to a variable extent. In the future, perioperative medicine will be fundamentally altered by the application of novel research technologies in proteomics to the clinical realm in the operating room, the intensive care unit, and the chronic pain clinic. The rapid improvement of techniques to identify changes in protein expression and function for diagnostics and the evaluation of therapeutic interventions is rapidly transforming our understanding of human physiology. It is important for the consultant anesthesiologist to be familiar with this emerging technology and its potential applications to the field as a whole.
The Technology
Our genome is complex, but our proteome—or all of the proteins coded for by genes—is still more complicated, more redundant, and more difficult to analyze and quantify. The need to connect the dots from genes to proteins is a major obstacle to the development of focused, mechanism-based therapeutics for a wide range of clinical diseases. There are approximately 30,000 human genes that are translated into as many as several million unique proteins. Proteomics represents the organized scientific effort to leverage high-throughput (i.e., processing large numbers of samples simultaneously) technologies and bioinformatics methods to generate a protein map, like that of a megacity mass transit system, that defines precisely the role of individual proteins in complex biochemical processes. Such a map, along with detailed structural information of individual proteins, could then be leveraged toward both the development of targeted molecular therapies for disease and the precise characterization of such complex physiologic states as unconsciousness from general anesthesia or prolonged critical illness (1).
By itself, proteomics is nothing new; rather, it represents a massive scale up of previous efforts and abilities. Understanding the fundamental principles of protein–protein interactions, enzyme catalysis, and related signal transduction pathways is now a central theme in nearly all biomedical research (2,3). Protein-based biologics, such as monoclonal antibodies and growth factors, comprise a growing percentage of novel drug therapies and diagnostic reagents currently under development. In 2002, nearly 35% of new drugs approved by the Food and Drug Administration in the United States were protein-based, and this trend continues today.
At first glance, the path from genetic code to functional protein appears elegant yet simple. Genes (DNA) are transcribed into messenger RNA (mRNA), which is subsequently translated (or expressed) into protein. Expressed protein performs a designated cellular function for a finite lifetime, and it is then targeted for breakdown and disposal by cellular mechanisms. It was once thought that this stepwise process was relatively straightforward. However, accumulated knowledge of cellular biochemistry, along with the massive data from the Human Genome Project, demonstrates that this process is actually incredibly complex and is modulated and redirected in multiple ways at every level on a continual basis. The transcription of the DNA of a single gene to mRNA is regulated by a host of proteins, as is the rate at which mRNA is metabolized or sequestered before it is translated into its protein product. It was initially thought that mRNA could be measured as a surrogate for protein activity; it is now understood that the quantitative relationship between them is nonlinear at best and frequently unreliable (4). For example, decreased mRNA levels may take minutes to weeks to influence protein activity, depending on the protein turnover time in a given cell, and copies of mRNA may not always be translated to their coded protein end products. Instead, it has become evident that proteins need to be studied and quantified directly.
The functions of proteins themselves are modified by chemical changes, dynamic folding states, and precise intracellular localization. The traditional notion of extracting a single protein from the cellular milieu and studying it in isolation has been accelerated by high-throughput technologies. To some extent, it has also been supplanted by methods that allow protein expression and function to be studied in the complex in vivo biochemical environment. This is the heart of the proteomics revolution, and it is the data coming from such studies that will shape clinical breakthroughs for the foreseeable future.
New technologies are enhancing our understanding of all areas of protein science (Table 1). Protein X-ray crystallography is a method in which small, painstakingly cultivated crystals of a protein are bombarded with X-rays, yielding a pattern of scatter that, interpreted by computer, produces a three-dimensional picture of the protein with a resolution on the order of a few Angstroms. These crystal structures can be critical to understanding function. It was, for example, the crystal structures of Human Immunodeficiency Virus (HIV) protease that led to the rapid development of HIV protease inhibitors for the treatment of HIV/AIDS. With the advent of robotic technology, supercomputers, and advanced molecular modeling systems, protein crystals are now grown and analyzed, and the resulting structures are crossed with possible drugs in high-throughput fashion at companies such as Structural GenomiX (San Diego, CA). The medical potential held in a structural database of all human proteins has recently led to the formation of the Protein Structure Initiative, sponsored by the National Institutes of Health, in which industry, government, and academia will bring unified resources to bear on this challenge (5,6). This approach may prove useful with respect to a diverse range of proteins. Of particular clinical interest are the liver cytochrome P450 enzymes responsible for drug metabolism. Recently, the first X-ray crystal structures of the human cytochrome P450 2C9 and 3A4 isoforms, complexed with clinically important medications (progesterone and warfarin), were determined (7,8) by Astex Technology (Cambridge, United Kingdom). Similarly, advances in automated patch-clamping technology are allowing high-throughput analysis of voltage-gated ion channels (9). Such channels represent an important component of the human proteome, constitute a target for many drugs (including inhaled anesthetics), and underlie a diverse range of pathological processes.
|
Unlike genetic material, such as DNA, proteins are inherently more complex and are not easily amenable to direct chemical synthesis, standardized purification, or efficient replication. The Human Genome Project was buttressed substantially by the ability to rapidly, reliably, and inexpensively synthesize strands of genetic material on modest scales through technologies such as the polymerase chain reaction. In contrast, proteins are structurally diverse, highly susceptible to their chemical and thermal environments, and typically undergo substantial yet simple chemical modification in vitro that dramatically alters their function. The most prominent example of this is phosphorylation of a protein involved in cellular signaling. The simple attachment of a phosphate group to a single amino acid side chain can be a sufficient stimulus to activate or shut down a major biochemical pathway. The cellular or plasma concentrations of such modified proteins can be astonishingly small. Purifying, or even identifying, individual proteins from this complex milieu was once a remarkable challenge and is now becoming approachable with new technologies.
Protein purification methods provide larger quantifies of desired products with increasing efficiency. Liquid chromatography separates proteins based on factors ranging from polarity to molecular weight. Advanced methods include the incorporation of specific binding moieties (such as antibodies) onto a resin to allow the separation of substrates of interest. Capillary electrophoresis separates molecules based on their mass-to-charge ratio, requires the application of only very small sample amounts of material, and is very useful for the detection of chemically modified proteins (10). A major development for protein separation has been the design of reliable reversed-phase chromatography systems, which can separate proteins according to their overall hydrophobicity. These systems are becoming increasingly automated and are able to handle both small volumes and large numbers of samples simultaneously. Different modalities are also being combined to produce mixtures with better component resolution; this includes more effective detection of low-abundance proteins and proteins with atypical physicochemical properties. Separation products can be directly subjected to inline mass spectrometric and bioinformatics analysis to yield rapid, simultaneous characterization of large numbers of proteins.
Traditionally, the standard method for separation of mixtures of proteins was gel electrophoresis in so-called two-dimensional gels. Such gels separate proteins that are run in a gel along two axes based on size and charge, respectively. Individual "spots" on a gel can then be extracted and studied by other methods. Two-dimensional gels can be tedious to execute, difficult to automate, and often impossible to standardize across laboratories (11). The gels are useful in that they separate proteins based on both size and charge, but they are much less effective in the separation of large hydrophobic, highly charged, and low-abundance proteins. Moreover, the separated proteins cannot be applied to analytic mass spectrometric analysis without substantial chemical pretreatment.
Chemical synthesis of large proteins, whereas feasible, is time consuming, environmentally harmful, and extraordinarily expensive. However, new approaches are developing that allow the use of other cells or organisms as bio-reactors to produce human protein in large quantities. For example, the fusion of a tumor cell with uncontrolled growth and synthetic potential with an antibody-producing B-cell resulted in the hybridoma (12). This is essentially a bio-reactor used to produce most fully human monoclonal antibodies. In similar fashion, plants, animals such as cattle, and even bacteria are being evaluated as potential reactors for the synthesis of other proteins of research and therapeutic interest. As these strategies evolve, ever larger quantities of diverse proteins will be available for research and therapy.
The availability of increasing numbers of purified proteins has naturally led to the application of DNA-based microarray technology to protein analysis. These microarrays are small silicon wafers onto which very small concentrations of biological material are attached. The material can then be screened en masse for an interaction with another DNA strand, small molecule, or protein. In a sense, hundreds or thousands of micro-experiments can be performed simultaneously. The resulting interactions are characterized by computer analysis.
Research using DNA-based technology includes exciting clinical applications such as GeneChips (Affymetrix, Santa Clara, CA) for screening of gene mutations associated with cancer or more subtle patient-specific polymorphisms associated with ion channel function and drug (cytochrome P450) metabolism. Whereas oligonucleotide microarrays for DNA and RNA have proved simple to manufacture, protein counterparts have required more effort. Proteins do not all share a common functional group to tether to the chip. Other problems include the effects of tethering on the three-dimensional structure of proteins and potentially the inability to function in a native capacity while restrained in space. These problems are gradually being overcome, and protein chips, as they are called, are being used regularly in the proteomic effort (13,14). These silicon wafers, widely used in the study of gene mutations, can have hundreds of individual proteins or diagnostic antibodies against plasma proteins attached to their surface (15). Methods for deposition and attachment of proteins as micro-spots have advanced considerably, as have strategies for the detection, amplification, and characterization of the resulting interactions (16). Protein microarray systems (Fig. 1) can be used to quantify small changes in protein levels that might correlate with a disease process, drug administration, or a metabolic derangement (18).
|
As a complement to microarray technology, highly sensitive mass spectrometers, novel labeling strategies, high-throughput liquid chromatography separation systems, and microarray technology provide a powerful armamentarium of tools for studying proteins in a variety of real and simulated environments (Fig. 2). In mass spectrometry (MS), the development of novel ionization and detection methods and their coupling to efficient online separation technologies and novel bioinformatics algorithms now allows rapid analysis of mixtures as complex as human plasma for differential protein composition. MS involves bombardment of the molecule of interest with laser energy to volatilize the compound and trigger the formation of intermediate species, such as ions or fragments, that can be subjected to analysis based on charge/mass ratios and diffusion rates in an electric field. Early methods of MS used high energy sources that either decomposed chemically sensitive biomolecules or were unable to volatilize large molecules and incorporated detection systems with insufficient resolution to handle large molecular weight intermediates in complex mixtures. New methods have been developed that have successfully circumvented these problems and have brought MS to the forefront in the analysis of complex protein mixtures. The methods of particular relevance include electrospray ionization, matrix-assisted laser desorption ionization time-of-flight (MALDI-ToF), surface-enhanced laser desorption time-of-flight (SELDI-ToF), and tandem MS/MS (Table 1).
|
The newer methods, such as MALDI, can analyze ata- or femtomolar quantities of material. They use gentler ionization methods to volatilize extremely small quantities of biological macromolecules that are embedded in a solid matrix that promotes the desorption process. The gentle ionization method allows large molecules such as proteins to be identified as the parent molecule rather than complex fragments. The addition of tandem MS/MS allows individual molecular peaks to be selected and bombarded with even more energy, which can, for example, decompose a protein into fragments from which the entire sequence can be gleaned. Eventually, as the technology evolves, such identification will be quantitative, and direct comparisons of protein concentrations in two different samples will be feasible. In such experiments, a sample of serum from a range of patients is injected into the spectrometer and generates a complex output of peaks (known as a protein fingerprint) corresponding to the masses and intensities of all of the protein components. With strict standardization protocols and the use of advanced computer analysis (see below), subtle differences in protein levels between healthy individuals and ill patients may identify diagnostic markers for the disease or possible therapeutic targets for drug development.
The analysis of samples extracted from blood or tissue provides information on the system at one point in time. In contrast, strategies for the identification of proteins in vivo, as they function in cells at native concentrations over a series of time points, will even further extend our ability to understand the complex interactions of protein signal transduction cascades (19,20). In one broadly applicable strategy, genetic material encoding fluorescent proteins is embedded into the code for genes of interest. When expressed, the resulting protein is coexpressed with the fluorescent marker and can be tracked through the cell with sophisticated optical imaging technology. Naturally occurring marine fluorescent proteins, notably green fluorescent protein, have now been modified and are available in a range of colors, sensitivities, and sizes. This diversity of both excitation and emission wavelengths allows for multiple proteins (colors) to be tracked simultaneously, and increased brightness (quantum yield) allows for the detection of fusion proteins by fluorescence scanners and microscopes at smaller concentrations.
There are several major pitfalls with this technology. They include the covalent modification of proteins under study with a relatively large molecular appendage (on the order of 27 kDa) that could alter the nature of native in vivo protein–protein interactions, interfere with correct protein folding, or attenuate enzymatic activity. Also, fluorescence of these proteins is inherently sensitive to native cellular conditions. For example, these proteins do not fluoresce under anaerobic conditions. Furthermore, the natural heterogeneity of protein expression and localization in individual cells makes the quantitative interpretation of fluorescent images intrinsically complex.
Fluorescence resonance energy transfer is a complementary approach that has advanced the application of fluorescence technology to functional proteomics (21). It relies on the direct transfer of energy from a small molecule in an excited state to an acceptor molecule that is within several nanometers of the donor. The resulting quenching of the emitted energy can be assayed and quantified by optical analysis. This approach is particularly useful for confirming the close proximity of two proteins in the cellular milieu. Such interactions can be monitored in real time, providing detailed insight into the functioning of the cellular machinery.
Miller and Cornish (22) have devised a methodology in which proteins of interest are fused to the enzyme dihydrofolate reductase. Expression of the protein is then detected by the addition of the small molecule substrate dihydrofolate to the cell. Another elegant strategy involves the incorporation of nonnatural fluorescent amino acids directly into the protein of interest by expression in cells containing modified transfer RNA species (22). These and related technologies offer the potential to perform proteomics analysis at the level of cells and tissues, where the physiologic effects of changes in protein expression and function are most clinically relevant.
Despite the aforementioned difficulties with mRNA as a surrogate marker for protein, direct inhibition of mRNA translation into proteins is another increasingly studied approach for modulating protein levels in the cell. For example, anti-sense mRNA (or short interfering RNA [siRNA]), which affects the cleavage and inactivation of complementary mRNA transcripts, has been investigated by companies such as Sirna (San Francisco, CA). These mRNA complements can inhibit protein activity in a direct fashion and thereby reveal aspects of individual protein function or potentially provide clinical treatment. A specific siRNA sequence can be synthesized in the laboratory to be exactly complementary to a target mRNA. Placing the two together in the cell might effectively halt synthesis of a deleterious or malfunctioning protein (23).
Most would agree that, at present, our experimental technologies are generating data at a rate and a level of complexity that exceeds our interpretive skills. An average MS experiment, for example, might generate 10,000 to 1,000,000 individual data points, with countless potential interactions between data points. The relative ease with which multivariable experiments can be preformed has also stimulated a shift from hypothesis-driven experimentation to data-driven research in which data mining to identify patterns dominates specific hypothesis testing.
Substantial resources are being brought to bear toward the development of software products and international databases for the reliable analysis of computer-generated proteomics data. An integral feature of such programs is the need to establish methods of standardization of data sets across experiments, technology platforms, and physiological/biochemical systems. For example, when a mixture of proteins is placed into a mass spectrometer, the concentrations of some proteins are so small that their signals are indistinguishable from the background noise. Moreover, computer-analysis software must piece together the data signals from thousands of protein fragments to generate a match with an intact native protein. The important point is to understand that a disease state, such as sepsis, may involve a change in the expression of just a few proteins among thousands. The major changes often occur in acute phase proteins, such as albumin and C-reactive protein, and platelets not specific to a disease. Moreover, the change in disease-specific protein levels may be only one or two orders of magnitude and may vary greatly over the course of an illness or even with the time of day. Data analysis thus occurs within a small dynamic range, further increasing the potential for erroneous data interpretation.
Numerous reports have highlighted the need for a systematic approach to identifying proteins and changes in protein expression from MS data. Recently, mass spectrometric data supporting a rapid serum test for biomarkers in ovarian cancer have been called into question based purely on the experimental and data analysis methods (24). The basic problem centers on how to extract meaningful quantitative data from series mass spectra containing thousands of peaks with wide-ranging intensities. Seemingly simple issues, such as determining what actually constitutes a peak as opposed to noise or artifact, are in reality quite complex and can greatly influence the experimental outcome. Baggerly (25) is a leader in developing reliable methods for analysis of protein MS data. His laboratory has introduced practical statistical methods for reliably extracting peaks, determining noise thresholds, and even simulating expected spectra using a virtual MS program.
Another approach to handling the vast array of data is the use of artificial neural networks (ANN). Such networks have been applied to genomic analysis of microarray data and are now under active study in proteomics. The basic concept is to program computers as neurons to mimic brain processing of data and to "learn" to recognize patterns and apply that knowledge to make generalizations and predictions based on the patterns identified in the raw data. The ANN technology is still in its infancy but has been applied to several MS studies in proteomics, such as those by Ball et al. (26) with human brain tumors, and is becoming more widely used in other clinical areas, such as the detection of myocardial infarction, breast cancer, and ovarian cancer (27). ANN does require programmers to define the "training set" without providing too much instruction that may hinder learning, and methods to perform this effectively are still in an early stage of development. A detailed discussion of bioinformatics resources and neural networks can be found in a topical review by Kepetanovic et al. (28) and in the web resources listing. Table 2 provides some useful and informative web-based educational resources for the technology covered in this section.
|
Clinical Applications
Proteomics studies carry the promise of providing highly translatable results and applications. This follows directly from the observations that protein biomarkers are growing in importance for the study of disease processes. Thus, polypeptides such as troponin, C-reactive protein, tumor necrosis factor, and brain natriuretic peptide have all been associated with cardiovascular and inflammatory diseases of widespread clinical importance. Indeed, the Food and Drug Administration has recently emphasized the role of protein biomarkers in clinical drug development as part of the "Critical Path" initiative (29). This initiative encourages the adoption of novel technologies to make clinical drug development both safer and more efficient; it includes an emphasis on the role of protein biomarkers as both surrogate end-points in difficult-to-study diseases and as predictors of drug toxicity and efficacy.
Differential protein expression in disease states, such as prostate, breast, and colorectal cancer, has already been demonstrated (30,31). For example, Liu et al. (30) analyzed protein profiles of digested prostrate specimens using SELDI MS and Ciphergen (Fremont, CA) ProteinChip® technologies. They found that prostate cancers with similar TNM classifications had similar protein expression profiles. Moreover, there was a downregulation of a specific enzyme, tissue inhibitor matrix metalloproteinase 1, in tumor samples compared with controls. This is consistent with a previous hypothesis regarding the role of matrix metalloproteinase enzymes in cancer. This approach is being extended to examine protein profiles in response to chemotherapy regimens and may help to predict metastatic potential and responsiveness to treatment. A new consortium of United States research institutions with significant federal funding through the National Cancer Institute/Food and Drug Administration Clinical Proteomics Program has recently been formed to focus on the identification of protein biomarkers for diagnostic and prognostic use in cancer (32,33).
Likewise, for many years, there has been tremendous interest in protein profiling in cardiovascular disease states. Most recently, serum levels of brain natriuretic peptide have become useful in the evaluation of congestive heart failure exacerbations. In a recent paper, Anderson (34) notes that there are presently nearly 200 secreted proteins that are of potential clinical interest in the evaluation of cardiovascular disease states. He notes that in contrast to the commonly used shotgun strategy, a more tractable approach to functional clinical proteomics is to begin with a small subset of candidate proteins on which to base further studies and hypotheses.
Proteomics methods are also beginning to be applied in the study of hepatic and pulmonary disease. As Sepper and Prikk (35) illustrate, disruption of the complex balance of matrix metalloproteinases, matrix metalloproteinase inhibitors, and matrix proteins is a recurring theme in chronic inflammatory lung disease. A better understanding of the dynamic function and interaction of these proteins may help to direct future therapies. However, the sheer number of potential models requires an integrated, broad-based approach that has been lacking in traditional research protocols. Similarly, Durr et al. (36) reported on the proteome profile of lung endothelial tissue in vivo, presenting results that shed light on the biochemical function of these cells and that will have implications for a broad range of disease processes.
The use of proteomics tools to study enzyme function has stimulated a new field of metabolite analysis referred to as metabolomics. The application of nuclear magnetic resonance spectroscopy and newer MS methods to profiling of serum metabolites has yielded interesting results in the characterization of enzymes with previously unknown function (37). For example, Allen et al. (38) used high-throughput electrospray ionization MS analysis of yeast culture extracts to detect global changes in metabolite profiles that result from deletion of cellular enzymes. Such strategies could be very useful for detecting global changes in cellular metabolism under conditions such as general anesthesia.
Proteomics and Anesthesia
Despite the vast potential for laboratory and clinical applications of proteomics technology in anesthesia, published data have been limited. Xi et al. (39) applied a combination of two-dimensional gel electrophoresis separation methods, 14C-halothane photolabeling, and MALDI-ToF MS to the identification of protein-binding sites for this inhaled anesthetic in the rat brain. A group of potential targets that included mitochondrial proteins and voltage-gated ion channels was identified. These data, consistent with the hypothesized modulation of oxidative phosphorylation by inhaled anesthetics (40,41), provide a basis for more focused proteomic studies of anesthetic binding sites. However, the limited ability of the techniques used to identify low-abundance proteins, which may represent clinically important binding sites, reinforces the need for continued investigation of these promising early findings.
Fütterer et al. (42) showed that desflurane anesthesia alters the levels of protein expression in the rat brain. The brain tissues of desflurane-exposed and desflurane-naïve rats were analyzed using a combined two-dimensional gel electrophoresis and MS approach. Changes in the expression of a number of proteins were detected. These included proteins involved in cellular trafficking, signal transduction, and mitochondrial function. Interestingly, the influence of anesthesia on protein expression levels seems to continue for at least 72 h after anesthetic exposure, which reinforces the notion that the physiological effects of general anesthesia outlast the immediate operative period.
The concept of phenotype-specific drug delivery is gradually becoming a reality in clinical practice. Mutation-specific chemotherapy is used routinely for the treatment of cancer and a variety of heritable metabolic derangements. In contrast, anesthetic drugs are largely administered in the absence of detailed mechanistic information and dosed on the basis of average distribution curves, accumulated clinical experience, and a continuous assessment of patient vital signs.
Proteomics holds the key to a deeper understanding of patient-specific drug response and the related global pharmacodynamics of drug action. Pain management is an area of both major clinical challenge and ballooning scientific knowledge for anesthesiologists. As the number of receptors, chemical signals, and modifiable pathways grow, so do the complexities of potential interactions. For example, clinicians may one day have available individual patient opioid receptor profiles, including information on the density of receptor subtypes and the response of the receptors to various therapeutics. Such information would be invaluable for the management of both acute perioperative pain and also chronic pain syndromes. The detailed proteomic investigation of opioid receptors, N-methyl-d-aspartate, and other receptors in the laboratory may also lead to the identification of novel subtypes and specific mutations in patients with pain syndromes or resistance to opioid-based analgesia. As Birch et al. (43) explain, the detailed study of ion channels has already added substantial insight to our understanding of pain, and the application of novel high-throughput technologies to the study of ion channel proteins involved in pain processing holds great promise.
In the approaching proteomics era, preoperative evaluation and risk stratification will move beyond chest radiographs, electrocardiography, and physical examination. Blood samples will likely be assayed for a host of inflammatory protein markers that may aid in the assessment of cardiopulmonary function and preoperative risk stratification. Similarly, preoperative testing could one day include a proteomic analysis of cytochrome P450 isoforms and serum esterase enzyme profiles. Such information would help to guide the dosing of a range of medications, notably nondepolarizing neuromuscular blockers, that rely extensively on hepatic or plasma metabolism for clearance. This information would also aid in the prediction and avoidance of detrimental perioperative drug interactions. Perioperative morbidity, including renal failure and myocardial infarction, is difficult to assess with current clinical definitions and diagnostic tools. As serum protein fingerprinting becomes mainstream, patterns of troponin, creatine kinase-MB fractions, and other markers of cardiac or renal function will be applied in this area, as well.
In the intensive care unit, real-time proteomic analysis of patients with sepsis may allow rapid subclassification of the syndrome into variants that may better predict responsiveness to fluid resuscitation, IV steroids, activated protein C, antitumor necrosis factor drugs, or specific antibiotics. Eli Lilly (Indianapolis, IN) has been a leader in the application of proteomics to clinical disease and has worked to create the IN Proteomics Consortium to test new proteomics technology directly in drug development. Lilly has been especially focused on the study of protein expression in sepsis and has identified previously unknown biomarkers, such as activated protein C, in that condition (44). Similarly, protein expression profiling in acute respiratory distress syndrome or endocrinopathies of prolonged critical illness may yield information to aid in the diagnosis and classification of the disorders and guide mechanism-based management.
A microchip containing a large portion of the yeast proteome is now commercially available (Invitrogen, Carlsbad, CA) against which drug binding can be assayed. Fifty percent of yeast proteins share substantial sequence homology with their human counterparts. Potentially, any drugs from sevoflurane to dexmedetomidine could be incubated with these proteins to investigate potential binding sites. It is becoming clear that inhaled anesthetics work at a diverse array of target sites (45–47). Screening of inhaled anesthetics against large pools of both soluble and membrane-associated proteins and lipids holds promise for the identification of specific sites of action.
There can be little doubt that the global effects of inhaled and IV anesthetics, so widely used today, will alter protein expression levels in unexpected and potentially modifiable ways. In the laboratory setting, proteomic technologies offer tremendous promise of a better understanding of the mechanisms of anesthetic action as well as a way to explore the hitherto undetectable physiological changes that accompany a general anesthetic. Using a genomics chip approach, Sergeev et al. (48) have recently reported on the effects of ischemia and anesthetic preconditioning on the level of gene expression in hearts. Their results, if extended to study the protein function of the identified genes, demonstrate an important application of these technologies to the study of the global effects of surgery and anesthesia on the biochemical milieu. Such modifications are evidenced by dynamic changes in expression across a broad range of protein classes, including heat shock proteins and hypoxia-inducible factor-related proteins.
Hirota et al. (49) presented data demonstrating that halothane blocks protein expression in response to hypoxia that is mediated by the transcription regulator hypoxia inducible factor-1. In their study, ischemia was induced with a 110-min period of no perfusion to the beating heart. Although a precise measure of Po2 was not reported, hypoxia inducible factor-1 is part of a complex cellular system for the modulation of cell function in the presence of low oxygen tensions (Po2 approximately 40 mm Hg) (50). Deciphering the network of proteins involved in this response and the concomitant effects of exogenous anesthetics would be of great clinical value and will require the application of proteomics strategies.
There are applications for proteomics technologies in transfusion medicine as it relates to anesthesiology. As Reddy and Perrotta (51) explain, detailed knowledge of the plasma proteome could be applied to the more precise cross-matching and screening of banked blood products. A more comprehensive understanding of protein function in platelets and other coagulation factors could have a profound impact on the perioperative management of patients undergoing cardiopulmonary bypass, on those with unrecognized coagulopathies, or in the management of antiplatelet medications in high-risk patients. Toward this end, the National Heart, Lung and Blood Institute in the United States has recently announced research support for proteomic projects to directly study protein function in platelets (52).
|
Conclusions |
|---|
|
TopAbstractIntroductionConclusionsReferences |
|---|
Anesthesiology still faces fundamental questions concerning nearly every realm, from mechanisms of drug action to perioperative management for the reduction of cardiovascular, pulmonary, renal and neurologic injury. The application of proteomics methodology to these pressing questions is almost certain to improve the scientific foundations of the discipline and to enhance patient care.
|
Footnotes |
|---|
Supported, in part, by NIH GM55876 and GM65218.
|
References |
|---|
|
TopAbstractIntroductionConclusionsReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
