Clinical Proteomics: Revolutionizing Disease Detection and Patient

the dynamic changes that give rise to cascades of new diagnostic biomarkers. .... Amy VanMeter , Michele Signore , Mariaelena Pierobon , Virginia ...
7 downloads 0 Views 345KB Size
Download PDF
Clinical Proteomics: Revolutionizing Disease Detection and Patient Tailoring Therapy Emanuel Petricoin,*,† Julia Wulfkuhle,‡ Virginia Espina,‡ and Lance A. Liotta*,‡ NCI/FDA Clinical Proteomics Program, Office of Cell and Gene Therapy, Center for Biologics Evaluation and Research, Food and Drug Administration, Bethesda, Maryland 20892, and NCI/FDA Clinical Proteomics Program, Laboratory of Pathology, Center for Cancer Research, National Cancer Institute, Bethesda, Maryland 20892 Received January 14, 2004

The evolving discipline of Clinical Proteomics is more than simply describing and enumerating the systematic changes in the protein constituency of a cell, or just generating lists of proteins that increase or decrease in expression as a cause or consequence of disease. Clinical applications of proteomics involve the use of proteomic technologies at the bedside with the ultimate goal to characterize the information flow through the intra- and extracellular molecular protein networks that interconnect organ and circulatory systems together. These networks are both new targets for therapeutics themselves as well as underpin the dynamic changes that give rise to cascades of new diagnostic biomarkers. The analysis of human cancer can be used as a model for how clinical proteomics is having an impact at the bedside for early detection, rational therapeutic targeting, and patient-tailored therapy. Keywords: proteomics • pharmacoproteomics • patterns • mass spectrometry • protein microarrays

Introduction The post-genome era of molecular medicine is rapidly moving beyond transcriptomics, gene lists, and functional genomics to proteomics. The function and inter-linking of proteins is directly linked contextually to the cellular, tissue, and physiological microenvironment. The protein-protein interactions that drive complex biological processes can be characterized as a fluctuating information flow within the cell, and throughout the organism through protein pathways and the cellular protein “circuitry”1-6 The deranged molecular networks in cancer are not confined to the diseased cell, but extend out to the microenvironment of the tumor-host interface, the surrounding stromal and vascular compartments, and outward to the circulation macroenvironment.7 The recognition that cancer is a product of the proteomic tissue microenvironment has important clinical implications from both an early detection and therapeutic targeting point of view. The tissue microenvironment can spawn entirely new biomarker cascades that are amplified by subtle changes at the earliest times of tumor growth and invasion. The exchange of information in the communication linkages at the invasion interface can give rise to changes that are reflected in specific alterations of the proteome of the circulation. * To whom correspondence should be addressed. Dr. Emanuel F. Petricoin, CBER/FDA, Bldg. 29A/2D12, 8800 Rockville Pike, Bethesda, MD 20892. Tel: (301) 827-1753. Fax: (301) 480-5005. E-mail: [email protected]. Lance A. Liotta, CCR/NCI/NIH, 10 Center Drive, Bethesda, MD 20892. Tel: (301) 496-2035. E-mail: [email protected]. † NCI/FDA Clinical Proteomics Program, Office of Cell and Gene Therapy, Center for Biologics Evaluation and Research, Food and Drug Administration. ‡ NCI/FDA Clinical Proteomics Program, Laboratory of Pathology, Center for Cancer Research, National Cancer Institute. 10.1021/pr049972m CCC: $27.50

 2004 American Chemical Society

The view of individual therapeutic targets can be reoriented to that of a rational targeting of the entire deranged molecular network, extending both inside and outside the cancer cell. Ultimately, targeting, response assessment, and therapeutic monitoring will be individualized, and will reflect the subtle pre- and post-therapy changes at the proteomic level as well as the protein signaling cascade systems between individuals. The ability to visualize these interconnections both inside and outside a cell could have a profound effect on how we view biology, and can enable the realization of the recent emphasis on personalized combinatorial molecular medicine. We can use cancer as a case study where the evaluation and implementation of new proteomic tools is occurring at the bedside.

Protoemic Technologies: Early Diagnosis Sadly, despite the best advances in technology, imaging, and genomic information, cancer today is most often diagnosed and treated when it is too late and metastasis has already occurred. At this time, therapeutic modalities are very limited in their success and most of treatment is spent providing palliative quality-of-life therapy. Detecting cancers when they are at their earliest stages could result in better clinical outcomes. It is selfevident that the best way to cure cancer is to detect it before it has metastasized. An example of this is ovarian carcinoma. Today, more than two-thirds of ovarian cancers are detected at later stages (III and IV), when the ovarian cancer cells have spread and have disseminated throughout the peritoneal cavity.8 Even though the disease at this stage is advanced, it rarely produces specific or diagnostic symptomssbloating, crampiness, and fatigue are cited as the most frequent observations. As a result, ovarian cancer is usually treated when it is at an advanced stage, and the 5-year survival is 35-40% for Journal of Proteome Research 2004, 3, 209-217

209

Published on Web 03/02/2004

reviews late stage patients even in the face of cutting-edge treatment modalities.9 Conversely, when ovarian cancer is detected early (stage I), simple conventional therapy produces a 5-year survival rate approaching 95%.9-13 Therefore, because of this great clinical need, a major focus of marker discovery has been for ovarian cancer and high screening indications. For greatest clinical use, a biomarker should be measurable in an easily accessible body fluid such as serum, urine, or saliva. Because these fluids are a protein-rich information source that effectively contains what the circulation has encountered on its journey throughout the body and tissue perfusion, proteomic technologies may have the most impact in biomarker discovery. Unfortunately, past biomarker discovery efforts have centered on laborious approaches looking for the elusive single overexpressed protein in blood.14-19 Since there are tens to perhaps hundreds of thousands of intact, modified, and cleaved protein isoforms in the human serum proteome, most of which have not been elucidated, finding the single biomarker is like searching for a needle in a haystack, requiring the separation and identification of each protein biomarker. The very small number of newly approved biomarkers19 is an unfortunate reflection of the inability and failure of hypothesis-driven and low-throughput approaches to deliver clinically useful biomarkers. A major source of this problem stems from our lack of basic knowledge about the proteomic components of serum and plasma. Only recently have new approaches begun to identify the proteomic component of serum.20 Moreover, it is highly likely that a single marker for a cancer does not exists cancer is not a process of infection by a foreign body, which can be identified by unique proteins not found normally in the body. It is a disease of deranged host cells, which do not produce some entirely “new” molecule. Clinical applications will be eventually applied to a human population constituted by vast heterogeneity not only in their respective proteomes, but also in the underlying disease process itself. Thus, it seems reasonable that the presence of cancer, with high sensitivity and specificity, will be detected by multiplexed panels of clinical tests that measure modified and clipped/cleaved host and tumor-derived proteins, produced as a consequence of aberrant cellular function and cellular interactions.

Mulitplexed Proteomic Tools: Proteomic Pattern Diagnostics Unlike traditional biomarker approaches, mass spectrometry, provides for an extremely rapid and potentially high-throughput method for biomarker discovery. It would not be unreasonable to process tens of thousands of samples per week per instrument. The platform can instantaneously distinguish protein isoforms that differ by extremely small changes in molecular weight, providing a fingerprint of a component of the proteomic content within the specimen. The complexity of that fingerprint will change depending on the resolution and mass accuracy of the instrument itself. Of course, regulatory review and approval of pattern recognition-based endpoints will be treated on a case-by-case basis as regulatory agencies and industry gain expertise with multiplexed assays and high dimensional data emanating from gene transcript microarrays, protein microarrays, NMR, and mass spectral fingerprints.21 Indeed, recent attempts to employ mass spectroscopy for the identification of biomarkers for cancer have been very promising.22-28 Unlike past attempts that start with a known single marker candidate, proteomic pattern analysis begins with high dimensional data, usually produced by high-throughput 210

Journal of Proteome Research • Vol. 3, No. 2, 2004

Petricoin et al.

mass spectrometry. This method attempts, without bias, to identify patterns of low molecular weight biomarkers as ion peak features within the spectra as the diagnostic itself. Moreover, this strategy is not dependent on the poor cycle time that plagues traditional biomarker research in which the discovery platform and clinical assay system are unlinked: (a) discovery of a differentially expressed gene/protein; (b) generation of specific antibodies; (c) validation of antibodies; (d) generation of immunoassay platform; (e) validation of platform; (f) preclinical validation of marker; and (g) clinical validation. Mass spectral pattern diagnostics couples the discovery platform and clinical platform together (the mass spectrometer) and evaluates millions of combinations of potential biomarkers at once on large clinical sets. Thus, the patterns that emerge can be immediately validated on blinded statistically significant study sets. Some investigators29 have voiced concern about a process that does not require protein identity before clinical implementation, pointing to the identification as a necessity as it will provide clues to the biology, or serve as targets. This concern is somewhat unfounded as the oncology field is filled with known and fully sequenced biomarkers which have nothing to do with the disease process itself or provide clues to the biology (e.g., PSA), and biomarkers which have been measured for years without even knowing the underlying identity and amino acid sequence (e.g., CA-125). Thus, the myth that biomarker identity is needed for clinical use is not supported by the historical context. Serum Proteomic Pattern Diagnostics: Producing the Mass Spectra. While investigators have used a variety of different bioinformatic algorithms for pattern discovery, the most common analytical platform is comprised of a Protein Chip Biomarker System-II [PBS-II, a low-resolution time-of-flight (TOF) mass spectrometer (MS)]. Herein, samples are ionized by surface-enhanced laser desorption/ionization (SELDI), a Protein Chip array-based chromatographic retention technology that allows for direct mass spectrometric analysis of analytes retained on the array (Figure 1). Only a subset of the proteins in the serum bind to the chromatographic surface of the chip, and the unbound proteins are washed away. The bait region containing individual captured serum protein samples is overlaid with a coating of an organic acid matrix (e.g. a-cyano5-hydroxycinnamic acid) which crystallizes and then the entire chip is inserted into a vacuum chamber and a laser beam is fired at each spot. The organic acid matrix serves as an energy transfer medium for protein ionization, whereby the kinetic energy from the laser causes protein desorption/ionization.30 The mass to charge value of each ion is estimated from the time it takes for the launched ion to reach the electrode; small ions travel faster. Consequently, the spectrum provides a “timeof-flight” (TOF) signature of ions ordered by size. Recently, this concept has been extended to a high-resolution MS as it has been found for ovarian cancer detection that higher resolution MS data generates diagnostic models possessing higher sensitivities and specificities as a result of both the increased number of peaks seen and the much better between and within run machine reproducibility.23 Moreover, the spectral resolution of the lower resolution instrumentation may not be able to separate specific ions that are close in mass/charge and which can coalesce multiple specific discreet ions into a single peak. Of course, whether low-resolution or high-resolution mass spectrometry is used as a clinical diagnostic platform remains to be seen, as more comparison studies are required. Clinical utility is not just predicated based on clinical performance. The

Clinical Proteomics

reviews

Figure 1. Mass spectrometry as a diagnostic tool. Surface enhanced laser desorption/ionization time-of-flight (SELDI-TOF) mass spectrometry is one type of proteomic analytical tool and is a class of mass spectroscopy instrument useful in high throughput proteomic fingerprinting of serum. Depending on the surface chemistry used, (WCX2 ) weak cation exchange surface; SAX2 ) strong anion exchange surface; IMAC3 ) immobilized metal affinity surface) a subset of the proteins in the sample bind to the surface of the chip with unbound proteins washed off after incubation. The bound proteins are treated with a MALDI matrix, washed and dried. The chip, containing multiple patient samples, is inserted into high (ABI Qstar) or low (Ciphergen PBS) resolution mass spectrometers and analyzed by laser desorption/ioization. The time-of-flight (TOF) of the ion prior to detection by an electrode is a measure of the mass to charge (M/Z) value of the ion, with most ions being singly charged. The mass spectra can then be analyzed by various pattern recognition software to discover potential diagnostic differences based not on one molecule, but on a pattern of multiple decreases and increases in ion amplitudes.

entire process will need to be evaluated for each step of the process: sample handling, archiving, database and processing standard operating procedures, sample application and robotic handling procedures, mass spectrometry cGMP and ISO9001 performance, protein chip and/or MALDI plate performance, software validation, and database management. The high-resolution mass spectrometer used in recent proteomic pattern based studies employs a hybrid quadrupole time-of-flight mass spectrometer (QSTAR pulsar i, Applied Biosystems Inc., Framingham, Massachusetts) fitted with a ProteinChip array interface (Ciphergen Biosystems Inc). As a point or analytical comparison, the Qq-TOF MS (routine resolution ≈ 8000) can completely resolve species differing in m/z of only 0.375 (e.g., at m/z 3000) whereas complete resolution of species with the Ciphergen PBS-II TOF MS (routine resolution ≈ 150) is only possible for species that differ by m/z of 20 (Figure 2). In a clinical setting where a pattern test may be eventually employed as a diagnostic, it will be important to determine overall spectral quality and develop spectral release specifications such that variances introduced into the process can be

evaluated and monitored. Day-to-day, lot-to-lot, and machineto-machine variances brought in from sample handling/storage and shipping conditions will need to be evaluated and understood, as well as the mass spectrometer itself. To that end, we routinely employ a pooled reference standard sample, obtained from NIST (SRM-1951A) which is randomly applied to one spot on each protein array as a quality control for overall process integrity, sample preparation, and mass spectrometer function. Additionally, for spectral quality control, quality assurance and spectral release specification, all spectra are subjected to plotting by total ion current (total record count), average/mean and standard deviation of amplitude, chi-square, and t-test analysis of each ion or bin, and quartile plotting measures using JMP (SAS Institute, Cary, NC) software as well as stored procedures that we developed in-house, prior to any pattern discovery. Process measures are checked by analyzing the statistical plots of the NIST serum reference standard, and spectra that fail statistical checks for homogeneity are eliminated from in-depth modeling and analysis This type of upfront analysis is critical so that it is possible to compare the total Journal of Proteome Research • Vol. 3, No. 2, 2004 211

reviews

Figure 2. Comparison between low resolution and high-resolution SELDI-TOF mass spectra. Spectra from the same weak cation exchange chip (queried at the same spot on the same chip) were generated on either a PBS IIc (low resolution instrument Ciphergen Biosystems, Inc.) (Panel A) or on a QSTAR Pulsar i highresolution instrument, (Applied Biosystems Inc., Framingham, Massachusetts) (Panel B).

analytical variance obtained for the constant NIST reference sample with the variance of the clinical sample populations The total variance of the reference sample should be no less than that for the clinical specimens. ProteinChip arrays (Ciphergen Biosystems Inc.) should be processed in parallel using robotic liquid handling to minimize variance (e.g., a Biomek Laboratory workstation (BeckmanCoulter)) and modified to make use of a ProteinChip array bioprocessor (Ciphergen Biosystems Inc.). The bioprocessor holds 12 ProteinChips, each having 8 chromatographic “spots”, allowing 96 samples to be processed in parallel. It is extremely important that the matrix is also applied using a liquid robotic handling station (e.g., Genesis Freedom 200, TECAN; Research Triangle Park, NC) as spectral quality is highly dependent on matrix deposition, drying and crystallization. Serum Proteomic Pattern Diagnostics: Uncovering the Pattern Classifiers. A typical low resolution SELDI-TOF proteomic profile will have up to 15 500 data points that comprise the recordings of data between 500 and 20 000 m/z, with a high-resolution mass spectrometer generating over 1 000 000 data points. A multitude of downstream pattern recognition systems exist, and all may show very good reliability at detecting and discovering sets of classifying ion features. We begin our own mass spectral proteomic pattern analysis by first exporting the raw data file generated from the high resolution QSTAR mass spectra into tab-delimited files that generate approximately 350 000 data points per spectrum. The spectra is then binned using a function of 400 parts per million (ppm) such that all data files possess identical m/z values (e.g., the m/z bin sizes linearly increase from 0.28 at m/z 700 to 4.75 at m/z 12 000). This binning process actually condenses the number of data points from 350 000 to exactly 7084 points per sample, and by a ppm binning function the m/z range of the bins gradually increases as a function of the resolution capacity of the machine. The 400 ppm binning function was based on a value obtained by a 10 times the estimate of what the mass drift of the Qq-TOF machine routinely obtains by external and internal calibration results (5-40 ppm)- as a conservative drift bracket. The data are then normalized (necessary since mass spectrometry is inherently nonquantitative) and then randomly separated into equal groups for training and testing. Data 212

Journal of Proteome Research • Vol. 3, No. 2, 2004

Petricoin et al.

normalization is an important element of pattern recognition so as to ensure a commonality in the spectra itself and assess potential for bias (e.g., introduced by protein chip quality, mass spectrometer instrumentation, and operator variance, sample collection, sample handling, and storage) and which can effect overall spectral performance and introduce potential nondisease related artifact into the spectra. It is likely that different data normalization procedures will generate different ions selected, especially in a clustering algorithm where multiple ion features are used as the pattern. Since mass spectrometry is not inherently quantitative, scalar intensity changes may be apparent, yet the overall pattern may not changed. Normalization can be achieved by dividing the spectra by the total ion current, amplitude value sums or average. One way we typically normalize mass spectral data is by dividing the amplitudes at each M/Z value within any randomly generated pattern subset by the largest value within that subset using a genetic algorithmbase approach. In this way, differences in spectral quality that may emanate from biases such as in protein chip variance and not from the inherent disease process itself, can be minimized. Also, this method allows for low amplitude features to contribute substantially to the classification. Our spectra and experimental description are posted at the following site: http://ncifdaproteomics.com. Mass Spectrometry Based Diagnostics: A View to the Future. Mass spectrometry analysis of the low molecular weight range of the serum/plasma proteome is a rapidly emerging frontier for biomarker discovery and clinical diagnostics. Proteomic pattern diagnostics represents a new paradigm for disease detection and is very amenable to the high throughput word of clinical diagnostics. The analysis requires only a drop of blood and the mass spectra patterns obtained in less than 30 min. Mass spectrometry driven pattern analysis, in theory, may be applied to any biologic state. Using this approach, the pattern itself, independent of the identity of the proteins or peptides, is the discriminator, and may be clinically useful immediately before the underlying identities are eventually discerned. Depending on the identity of the signature ion, it may, or may not, be desirable, or even feasible to proceed directly to develop a serum immunoassay for the individual biomarker. This is because the ion amplitude of MALDI-TOF does not directly reflect the concentration of the given biomarker associated with the ion. Moreover, if the biomarker is the cleaved version of a larger protein, it may be difficult to generate antibodies that recognize the cleaved version and do not cross react with the parent species. A possibility exists to develop polyclonal antibodies as a bait, and that following binding, the entirety of the recognized entities, including the diagnostic fragment are eluted and analyzed via mass spectrometry. Mass spectroscopy platforms of the future, coupled to pattern recognition algorithms may become superior to antibody-based immunoassays.31,32 Mass spectrometry can generate complex proteomic spectra from an extremely small volume of blood in only a few secondssin effect sensing the presence of hundreds to thousands of events simultaneously almost instantaneously without the need to develop antibodies for each analyte. Mathematically, it should be obvious that a pattern of multiple biomarkers will contain a higher level of discriminatory information compared to a single biomarker alone, particularly for large heterogeneous patient populations. As evidence of the growing acceptance of this new paradigm,

reviews

Clinical Proteomics

Figure 3. Biomarker Amplification and Harvesting by Carrier Molecules. Low molecular weight peptide fragments, produced at the interface of the diseased cell and the tissue microenvironment permeate through the endothelial cell wall barrier and trickle into the circulation. Here, these fragments are immediately are bound with circulating high abundance carrier proteins such as albumin and protected from rapid kidney clearance. The sequestration of the low abundance biomarkers by the carrier protein pool over time results in the net effect of an enrichment and amplification of the biomarker fragments. In the future, harvesting nanoparticles, engineered with high affinity for binding, can be instilled into the collected body fluids or perhaps even injected directly into the circulation. These nanoparticles and their bound biomarkers can then be collected, filtered over engineered nanofilters and directly queried by highresolution mass spectrometry. A look up table, where the exact identities of each of the peaks will be compared against the accurate mass tag of each of the peaks within the spectra (e.g., through the use of FTICR-type systems) will soon enable the simultaneous identification of each entity within the pattern as well as the discovery of the diagnostic pattern itself.

large commercial reference laboratories have begun initiatives to explore mass spectroscopy proteomic patterns for routine diagnosis.33 Recent exciting discoveries indicate that a vast majority of these biomarkers exist in association with circulating high molecular mass carrier proteins.34,35 These findings now shift the focus of biomarker analysis from a serum-wide analysis to the just the carrier protein and its biomarker content. The discriminatory molecules are likely to be metabolic products, enzymatically generated fragments, and modified protein fragments. In fact, the most important biomarkers may be normal host proteins that are aberrantly clipped, modified or reduced in abundance. A pattern analysis approach takes into consideration the loss or gain of ions within the spectra without bias unlike past discovery based methods. Until now, conventional protocols for biomarker discovery discard the abundant “contaminating” high molecular mass proteins, to focus on the low mass range. Unfortunately, this procedure removes most of the important diagnostic biomarkers.34,35 We can now develop new tools, created at the intersection of proteomics and nanotechnology, whereby nanoharvesting agents can be instilled into the circulation (e.g., derivatized gold particles) or into the blood collection device to act as “molecular mops” that soak up and amplify biomarkers via accumulation (Figure 3). These nanoparticles, with their bound diagnostic cargo, can be directly analyzed by mass spectrometry and the low molecular weight and enriched biomarker signatures revealed. Coupling this method with ultrahigh-resolution mass spectrometry (e.g., Fourier transform ion cyclotron resonance mass spectrom-

etry36,37) will allow for rapid protein identification and diagnostic analysis all at the same time with the same machine.

Clinical Proteomic Tools For Patient Tailored Therapeutics Obviously, the best way to identify those patients will respond to a given therapy would be to know (prior to treatment) which, of the many potential proteins that could be chosen to target, were actually “in use” in each patient. Optimally, this would come from analysis through a limited amount of material taken from the patient through biopsy procurement or through serum analysis. In general, proteomic technologies have significant limitations when they are applied to limited quantities of tissue samples. Tools such as 2-D gels, isotope-coded affinity tagging (ICAT) multidimensional LC-MS platforms, and antibody arrays require relatively large numbers of cell equivalentssmany orders of magnitude greater than the quantity procured during a clinical biopsy.38-46 These clinical specimens may only contain a few hundred cells as the starting point for analysis. Unfortunately, since there is no direct PCRlike technology for proteomics, new micro-proteomic technologies that can employ these tiny microscopic amounts of cellular material need to be developed. A second limitation of many existing multiplexing proteomic technologies is the inability to analyze native protein samples. Because denaturation will break apart protein complexes, protein and erase 3-D protein conformation, these methods may not adequately probe the state of the cellular circuitry mediated by proteinprotein interactions. Journal of Proteome Research • Vol. 3, No. 2, 2004 213

reviews

Petricoin et al.

Figure 4. Reverse Phase Protein Arrays. Lysates are prepared from cultured cells or microdissected cells from tissue, and arrayed as a series of dilutions as the lysate itself is immobilized. The specific analyte molecule contained in the immobilized and denatured sample is then detected by a separate labeled probe (e.g., antibody) applied to the surface of the array. Probing the array with validated phospho-specific antibodies (e.g., phospho-ERK 1/2) allows for the detection of specific phosphorylated signaling components and kinase substrates. This array has demonstrated highly linear and very sensitive analyte detection and requires no upfront tagging or pre-fractionation.47-53

Protein microarrays represent the first new technology that can actually profile the state of a signaling pathway target even after the cell is lysed and the contents denatured.47-53 A new type of protein array, the reverse phase protein array, can utilize the very small numbers of human cells (Figure 4) taken from human biopsy specimens, and a protein lysate obtained through laser capture microdissection is arrayed onto nitrocellulose slides. Components of this method offer unique advantages over tissue arrays54 and other types of protein arrays55-58 including the fact that this approach can utilize denatured lysates. Antigen retrieval issues are not problematic, which is a limitation for tissue array analysis. Moreover, reverse phase arrays can reveal subtle quantitative differences in protein phosphorylation. Relative subtle expression differences can be analyzed since each sample is arrayed in a miniature dilution curve, providing an internal standard curve for quantitative and qualitative analysis. Since reverse-phase protein microarrays do not require direct tagging of the protein as a readout for the assay, dramatic improvement in reproducibility, sensitivity, and robustness of the assay over other techniques is achieved. A growing body of evidence is emerging which supports the concept that each patient’s cancer may have a unique complement of pathogenic molecular derangements, and there exists strong justification for the strategy to select from a menu of treatment choices, or treatment combinations, those which best match the individual tumor’s molecular profile.59-65 Molecular profiling using gene arrays have shown considerable potential to classify patient populations according to disease stage or 214

Journal of Proteome Research • Vol. 3, No. 2, 2004

survival outcome.66 However, transcript profiling, by itself, may provide an incomplete picture, because gene transcript level may bear no relationship to the phosphorylated or otherwise functional state of the encoded protein. Gene transcripts provide little information about protein-protein interactions and the state of the cellular circuitry, with this information being inferred by correlative bioinformatic approaches. Pharmacogenomics using gene arrays for molecular profiling to select the appropriate treatment strategy should include a direct proteomic pathway analysis of the biopsy material, and not be determined by correlative science alone. Common sense would dictate the use of proteomic tools when proteomic endpoints are what are being studied (e.g., protein kinase activity). Currently, cancer therapy has been directed at a single molecular target. Armed with the information about which signaling pathways are being used by the cancer cells and the surrounding stroma will allow us to build function interaction maps for each patient, and see how this information changes and fluxes as a consequence of the treatment choice not just within the tumor itself, but in the surrounding cellular milieu. Because these protein interactions are interdependent on each other, kinase activity at one location will affect effect other kinases and substrates within the circuit. We can now take advantage of this fact so that soon, we can imagine targeting an entire set of nodes all along the pathogenic signal pathway with a higher potential for efficacy with a lower toxicity67,68 (Figure 5). The phosphoproteome, profiled using reverse phase

reviews

Clinical Proteomics

and protein network analysis combined with genomic analysis using microscopic quantities of patient tissue material and/or serum proteomic pattern analysis. On the basis of proteomic and genomic portraits of the disease, an individualized selection of therapeutic combinations that best target the protein network will be selected and employed resulting in a paradigm shift in patient treatment and disease management.

References

Figure 5. Combinatorial therapeutics offer reduced toxicity and increased efficacy. A generic cartoon of a signal cascade is depicted whereby interconnections of kinases and their substrates along with their binding partners is shown. Targeting a single upstream node (upper) at a high dose of the drug is required for 90% reduction of pathway activity. In contrast, targeting a series of specifically selected interconnected and interdependent nodes within the cellular circuitry can theoretically achieve the same efficacy with a lower dose of each drug if used in combination (lower).

protein microarrays, will play a key role in personalized medicine as the aberrant function of protein kinases are often at the center of many diseases, including cancer.69-75 The new focus of narrowly focused molecular targeted therapeutics addresses this concept. STI-571 (Gleevec, imatinib mesylate) is a key example, as treatment with STI-571 targets the dominant activity of the c-Abl kinase protein, not the gene. Although the result of this proteomic circuit has a defective genetic underpinning through a well-characterized chromosomal translocation, the effect is that the deranged proteomic function results in the c-kit family of protein being turned on-, which then results in aberrant growth and pro-survival functions.76,77 Drug discovery efforts focusing on the development of small molecular weight compounds and biologics that can mitigate and modulate specific kinase activity is an intense area of focus for industry due to their key roles in cancer and biology.78-83

Concluding Remarks The current and future development of new clinical proteomic tools will have important translational applications for early detection, as a supplement to existing and co-evolving technological advances in diagnostic imaging, and provide a rational basis for patient tailored therapy. In the future, the oncologist can take advantage of this proteomic armamentarium. Cancer could be detected earlier, with greater specificity and sensitivity using mass spectrometry as the central clinical tool coupled with artificial intelligence based pattern recognition systems. The intersection of nanotechnology and proteomics could result in tools that can sense early changes in the underlying pathophysiology through the comprehensive analysis of LMW biomarker fragments. Once detected, each patient’s cancer will be profiled through phosphoproteomic

(1) Liotta, L.; Petricoin, E. Molecular profiling of human cancer. Nat. Rev. Genet. 2000, 1, 48-56. (2) Ideker, T.; Thorsson, V.; Ranish, J. A.; Christmas, R.; Buhler, J.; Eng, J. K.; Bumgarner, R.; Goodlett, D. R.; Aebersold, R.; Hood, L. Integrated genomic and proteomic analyses of a systematically perturbed metabolic network. Science 2001, 292, 929-934. (3) Schwikowski, B.; Uetz, P.; Fields, S. A network of protein-protein interactions in yeast. Nat. Biotechnol. 2000, 18, 1257-1261. (4) Legrain, P.; Jestin, J. L.; Schachter, V. From the analysis of protein complexes to proteome-wide linkage maps. Curr. Opin. Biotechnol. 2000, 11, 402-407. (5) Blume-Jensen, P.; Hunter, T. Oncogenic kinase signaling. Nature Insight. Cancer 2001, 411, 355-365. (6) Pawson, T. Protein modules and signaling networks. Nature 1995, 373, 573-580. (7) Liotta, L. A.; Kohn, E. C. The microenvironment of the tumourhost interface. Nature 2001, 411, 375-379. (8) Tumor Markers. In: Principles and Practice of Gynecologic Oncology; Hoskins, W. J., Perez, C. A., Young, R. C., Eds.; Lippincott, Williams & Wilkins: Philadelphia, PA, 2000, 165-182. (9) Wulfkuhle, J. D.; Liotta, L. A.; Petricoin, E. F. Early detection: Proteomic applications for the early detection of cancer. Nat. Rev. Cancer. 2003, Apr 3(4), 267-275. (10) Bast, R. C.; Klug, T. L.; St. John, E.; Jenison, E.; Niloff, J. M.; Lazarus, H.; Berkowitz, H. S.; Leavitt, T.; Griffiths, C. T.; Parker, L.; Zurawski, V. R.; Knapp, R. C. A radioimmunoassay using a monoclonal antibody to monitor the course of epithelial ovarian cancer. N. Engl. J. Med. 1983, 309, 883-887. (11) Menon, U.; Jacobs, I. J. Recent developments in ovarian cancer screening. Curr. Opin. Obstet. Gynecol. 2000, 12, 39-42. (12) Jacobs, I. J.; Skates, S. J.; MacDonald, N.; Menon, U.; Rosenthal, A. N.; Davies, A. P.; Woolas, R.; Jeyarajah, A. R.; Sibley, K.; Lowe, D. G.; Oram, D. H. Screening for ovarian cancer: a pilot randomized controlled trial. Lancet 1999, 353, 1207-1210. (13) Cohen, L. S.; Escobar, P. F.; Scharm, C.; Glimco, B.; Fishman, D. A. Three-dimensional power Doppler ultrasound improves the diagnostic accuracy for ovarian cancer prediction. Gynecol. Oncol. 2001, 82, 40-48. (14) Adam, B. L.; Vlahou, A.; Semmes, O. J.; Wright, G. L., Jr. Proteomic approaches to biomarker discovery in prostate and bladder cancers. Proteomics 2001, 1, 1264-1270. (15) Carter, D.; Douglass, J. F.; Cornellison, C. D.; Retter, M. W.; Johnson, J. C.; Bennington, A. A.; Fleming, T. P.; Reed, S. G.; Houghton, R. L.; Diamond, D. L.; Vedvick, T. S. Purification and characterization of the mammaglobin/lipophilin B complex, a promising diagnostic marker for breast cancer. Biochemistry 2002, 41, 6714-6722. (16) Rosty, C.; Christa, L.; Kuzdzal, S.; Baldwin, W. M.; Zahurak, M. L.; Carnot, F.; Chan, D. W.; Canto, M.; Lillemoe, K. D.; Cameron, J. L.; Yeo, C. J.; Hruban, R. H.; Goggins, M. Identification of hepatocarcinoma-intestine-pancreas/pancreatitis-associated protein I as a biomarker for pancreatic ductal adenocarcinoma by protein biochip technology. Cancer Res. 2002, 62, 1868-1875. (17) Xiao, Z.; Adam, B. L.; Cazares, L. H.; Clements, M. A.; Davis, J. W.; Schellhammer, P. F.; Dalmasso, E. A.; Wright, G. L., Jr. Quantitation of serum prostate-specific membrane antigen by a novel protein biochip immunoassay discriminates benign from malignant prostate disease. Cancer Res. 2001, 61, 6029-6033. (18) Kim, J. H.; Skates, S. J.; Uede, T.; Wong, K. K. K. K.; Schorge, J. O.; Feltmate, C. M.; Berkowitz, R. S.; Cramer, D. W.; Mok, S. C. Osteopontin as a potential diagnostic biomarker for ovarian cancer. JAMA 2002, 287, 1671-1679. (19) Anderson, N. L.; Anderson, N. G. The human plasma proteome: history, character, and diagnostic prospects. Mol. Cell Proteomics 2002, Nov 1(11), 845-67. (20) Tirumalai, R.; Chan, K.; Preito, D.; Issaq, H.; Conrads, T.; Veenstra, T. Characterization of the Low Molecular Weight Human Serum Proteome. Mol. Cell. Proteomics 2003, 2(10) 1096-1103.

Journal of Proteome Research • Vol. 3, No. 2, 2004 215

reviews (21) Petricoin, E. F., 3rd; Hackett, J. L.; Lesko, L. J.; Puri, R. K.; Gutman, S. I.; Chumakov, K.; Woodcock, J.; Feigal, D. W., Jr.; Zoon, K. C.; Sistare, F. D. Medical applications of microarray technologies: a regulatory science perspective. Nat. Genet. 2002, Dec 32(Suppl), 474-479. (22) Petricoin, E. F.; Ardekani, A. M.; Hitt, B. A.; Levine, P. J.; Fusaro, V. A.; Steinberg, S. M.; Mills, G. B.; Simone, C.; Fishman, D. A.; Kohn, E. C.; Liotta, L. A. Use of proteomic patterns in serum to identify ovarian cancer. Lancet 2002, 359, 572-577. (23) Conrads, T. P.; Zhou, M.; Petricoin, E. F., 3rd; Liotta, L.; Veenstra, T. D. Cancer diagnosis using proteomic patterns. Expert Rev. Mol. Diagn. 2003, Jul 3(4), 411-420. (24) Li, J.; Zhang, Z.; Rosenzweig, J.; Wang, Y. Y.; Chan, D. W. Proteomics and bioinformatics approaches for identification of serum biomarkers to detect breast cancer. Clin. Chem. 2002, Aug 48(8), 1296-1304. (25) Petricoin, E. F., 3rd; Ornstein, D. K.; Paweletz, C. P.; Ardekani, A.; Hackett, P. S.; Hitt, B. A.; Velassco, A.; Trucco, C.; Wiegand, L.; Wood, K.; Simone, C. B.; Levine, P. J.; Linehan, W. M.; EmmertBuck, M. R.; Steinberg, S. M.; Kohn, E. C.; Liotta, L. A. Serum proteomic patterns for detection of prostate cancer. J. Natl. Cancer Inst. 2002, Oct 16 94(20), 1576-1578. (26) Yanagisawa, K.; Shyr, Y.; Xu, B. J.; Massion, P. P.; Larsen, P. H.; White, B. C.; Roberts, J. R.; Edgerton, M.; Gonzalez, A.; Nadaf, S.; Moore, J. H.; Caprioli, R. M.; Carbone, D. P. Proteomic patterns of tumour subsets in nonsmall-cell lung cancer. Lancet. 2003 Aug 9 362(9382), 433-439. (27) Adam, B. L.; Qu, Y.; Davis, J. W.; Ward, M. D.; Clements, M. A.; Cazares, L. H.; Semmes, O. J.; Schellhammer, P. F.; Yasui, Y.; Feng Z.; Wright, G. L., Jr. Serum protein fingerprinting coupled with a pattern-matching algorithm distinguishes prostate cancer from benign prostate hyperplasia and healthy men. Cancer Res. 2002 Jul 1 62(13), 3609-3614. (28) Hingorani, S. R.; Emanuel F.; Petricoin, E. F., III; Maitra, A.; Rajapakse, V.; King, C.; Jacobetz, M. A.; Ross, S.; Conrads, T. P.; Veenstra, T. D.; Hitt, B. A.; Kawaguchi, Y.; Zhou, Y.; Johann, D.; Liotta, L. A.; Crawford, H. C.; Putt, M. E.; Jacks, T.; Konieczny, S. F.; Wright, C. E.; Hruban, R. E.; Lowry, A. M.; Tuveson D. A. Preinvasive and invasive ductal pancreatic cancer and its early detection in the mouse. Cancer Cell 2003, Dec 10, 6-21. (29) Diamandis, E. P. Point: Proteomic patterns in biological fluids: do they represent the future of cancer diagnostics? Clin. Chem. 2003, Aug 49(8), 1272-1275. (30) Hillenkamp, F.; Karas, M. Mass spectrometry of peptides and proteins by matrix-assisted ultraviolet laser desorption/ionization. Methods Enzymol. 1990, 193, 280-295. (31) Petricoin, E. F.; Liotta, L. A. Mass Spectrometry-based Diagnostics: The Upcoming Revolution in Disease Detection. Clin Chem. 2003, Apr 49(4), 533-534. (32) Petricoin, E., 3rd; Liotta, L. A. The Vision for a New Diagnostic Paradigm. Clin. Chem. 2003, Aug 49(8), 1276-1278. (33) http://www.questdiagnostics.com/; http://www.labcorp.com/ (34) Mehta, A.; Ross, S.; Lowenthal, M.; Fusaro, V.; Fishman, D.; Petricoin, E.; Liotta, L. Biomarker Amplification by Serum Carrier Protein Binding, Disease Markers 2003, 19(1) 1-10. (35) Liotta, L. A.; Ferrari, M.; Petricoin, E. Clinical proteomics: written in blood. Nature 2003, 425, 905. (36) Shen, Y.; Tolic, N.; Masselon, C.; Pasa-Tolic, L.; Camp, D. G., 2nd; Hixson, K. K.; Zhao, R.; Anderson, G. A.; Smith, R. D. Ultrasensitive proteomics using high-efficiency on-line micro-SPE-nanoLCnanoESI-MS and MS/MS. Anal. Chem. 2004, Jan 1 76(1), 144154. (37) Shen, Y.; Tolic, N.; Zhao, R.; Pasa-Tolic, L.; Li, L.; Berger, S. J.; Harkewicz, R.; Anderson, G. A.; Belov, M. E.; Smith, R. D. Highthroughput proteomics using high-efficiency multiple-capillary liquid chromatography with on-line high-performance ESI FTICR mass spectrometry. Anal. Chem. 2001, 73, 3011-3021. (38) Gorg, A.; Obermaier, C.; Boguth, G.; Harder, A.; Scheibe, B.; Wildgruber, R.; Weiss, W. The current state of two-dimensional electrophoresis with immobilized pH gradients. Electrophoresis 2000, 21, 1037-1053. (39) Hanash, S. M. Biomedical applications of two-dimensional electrophoresis using immobilized pH gradients: current status. Electrophoresis 2000, 21, 1202-1209. (40) Li, J.; Wang, C.; Kelly, J. F.; Harrison, D. J.; Thibault, P. Rapid and sensitive separation of trace level protein digests using microfabricated devices coupled to a quadrupolestime-of-flight mass spectrometer. Electrophoresis 2000, 21, 198-210. (41) Gygi, S. P.; et al. Quantitative analysis of complex protein mixtures using isotope coded affinity tags. Nature Biotechnol. 1999, 17, 994-999.

216

Journal of Proteome Research • Vol. 3, No. 2, 2004

Petricoin et al. (42) Washburn, M. P.; Wolters, D.; Yates, J. R. Large scale analysis of the yeast proteome by multidimensional protein identification technology. Nature Biotechnol. 2001, 19, 242-247. (43) Krutchinsky, A. N.; Kalkum, M.; Chait, B. T. Automatic identification of proteins with a MALDI-quadrupole ion trap mass spectrometer. Anal. Chem. 2001, 73, 5066-5077. (44) Washburn, M. P.; Ulaszek, R.; Deciu, C.; Schieltz, D. M.; Yates, J. R., 3rd Analysis of quantitative proteomic data generated via multidimensional protein identification technology. Anal. Chem. 2002, 74, 1650-1657. (45) Zhou, H.; Ranish, J. A.; Watts, J. D.; Aebersold, R. Quantitative proteome analysis by solid-phase isotope tagging and mass spectrometry. Nature Biotechnol. 2002, 20, 512-515. (46) Zhou, G.; Li, H.; DeCamp, D.; Chen, S.; Shu, H.; Gong, Y.; Flag, M.; Gillespie, J.; Hu, N.; Taylor, P.; Emmert Buck, M.; Liotta, L. A.; Petricoin, E. F., III; Zhao, Y. 2-D differential in-Gel electrophoresis for the identification of human esophageal squamous cell cancer specific protein markers. Mol. Cell. Proteomics 2002, 1, 117-123. (47) Paweletz, C. P.; Charboneau, L.; Bichsel, V. E.; Simone, N. L.; Chen, T.; Gillespie, J. W.; Emmert-Buck, M. R.; Roth, M. J.; Petricoin, E. F., III; Liotta, L. A. Reverse phase protein microarrays which capture disease progression show activation of pro-survival pathways at the cancer invasion front. Oncogene 2001, 20, 19811989. (48) Nishizuka, S.; Charboneau, L.; Young, L.; Major, S.; Reinhold, W. C.; Waltham, M.; Kouros-Mehr, H.; Bussey, K. J.; Lee, J. K.; Munson, P. J.; Petricoin, E. F., III; Liotta, L. A.; Weinstein, J. N. Diagnostic markers that distinguish colon and ovarian adenocarcinomas: Identification by genomic, proteomic, and tissue array profiling. Cancer Res. 2003, 63, 5173-5646. (49) Espina, Virginia; Mehta, Arpita I.; Winters, Mary E.; Wulfkuhle, Julia; Petricoin, Emanuel, III; Liotta, Lance. Protein microarrays: molecular profiling technologies for clinical specimens. Proteomics 2003, Nov 3(11), 2091-2100. (50) Grubb, R. L.; Calvert, V. S.; Paweletz, C. P.; Phillips, J. L.; Linehan, W. M.; Gillespie, J. W.; Emmert-Buck, M. R.; Liotta, L.; Petricoin, E. F. Signal pathway profiling of prostate cancer using reverse phase protein arrays. Proteomics 2003, Nov 3(11), 2142-2146. (51) Wulfkuhle, J. D.; Aquino, J. A.; Calvert, V. S.; Fishman, D. A.; Coukos, G.; Liotta, L. A.; Petricoin, E. F., III. Signal pathway profiling of ovarian cancer from human tissue specimens using reverse-phase microarrays. Proteomics 2003, Nov 3(11), 20852090. (52) Herrmann, P. C.; Gillespie, J. W.; Charboneau, L.; Bichsel, V. E.; Paweletz, C. P.; Calvert, V. S.; Kohn, E. C.; Emmert-Buck, M. R.; Liotta, L. A.; Petricoin, E. F., III Mitochondrial Proteome: Altered Cytochrome Oxidase Subunit Levels in Prostate Cancer. Proteomics 2003, 3 No. 9, 1801-1810. (53) Liotta, L. A.; Espina, V.; Mehta, A. I.; Calvert, V.; Rosenblatt, K.; Geho, D.; Munson, P. J.; Young, L.; Wulfkuhle J.; Petricoin, E. F. Protein microarrays: Meeting analytical challenges for clinical applications. Cancer Cell. 2003, Apr 3(4), 317-25. (54) Torhorst, J.; Bucher, C.; Kononen, J.; Haas, P.; Zuber, M.; Kochli, O. R.; Mross, F.; Dieterich, H.; Moch, H.; Mihatsch, M.; Kallioniemi, O. P.; Sauter, G. Tissue microarrays for rapid linking of molecular changes to clinical endpoints. Am. J. Pathol. 2001, 159, 2249-2256. (55) Sreekumar, A.; Nyati, M. K.; Varambally, S.; Barrette, T. R.; Ghosh, D.; Lawrence, T. S.; Chinnaiyan, A. M. Profiling of cancer cells using protein microarrays: discovery of novel radiation-regulated proteins. Cancer Res. 2001, 61, 7585-7593. (56) MacBeath, G. Proteomics comes to the surface. Nat. Biotechnol. 2001, 19, 828-829. (57) Walter, G.; Bussow, K.; Lueking, A.; Glokler, J. High-throughput protein arrays: prospects for molecular diagnostics. Trends Mol. Med. 2002, 8, 250-253. (58) Kuruvilla, F. G.; Shamji, A. F.; Sternson, S. M.; Hergenrother, P. J.; Schreiber, S. L. Dissecting glucose signaling with diversityoriented synthesis and small-molecule microarrays. Nature 2002, 416, 653-657. (59) Karpati, G.; Li, H.; Nalbantoglu, J. Molecular therapy for glioblastoma. Curr. Opin. Mol. Ther. 1999, 1, 545-552. (60) Brown, C. K.; Kirkwood, J. M. Targeted therapy for malignant melanoma. Curr. Oncol. Rep. 2001, 3, 344-352. (61) Frankel, A. E.; Sievers, E. L.; Scheinberg, D. A. Cell surface receptor-targeted therapy of acute myeloid leukemia: a review. Cancer Biother. Radiopharm. 2000, 15, 459-476. (62) Cheng, J. D.; Rieger, P. T.; von Mehren, M.; Adams, G. P.; Weiner, L. M. Recent advances in immunotherapy and monoclonal antibody treatment of cancer. Semin. Oncol. Nurs. 2000, 16, 2-12.

reviews

Clinical Proteomics (63) Gasparini, G.; Gion, M. Molecular-targeted anticancer therapy: challenges related to study design and choice of proper endpoints. Cancer J. Sci. Am. 2000, 6, 117-131. (64) Cimoli, G.; Bagnasco, L.; Pescarolo, M. P.; Avignolo, C.; Melchiori, A.; Pasa, S.; Biasotti, B.; Taningher, M.; Parodi, S. Signaling proteins as innovative targets for antineoplastic therapy: our experience with the signaling protein c-myc. Tumorigenesis 2001, 87, S20-S23. (65) Kolonin, M.; Pasqualini, R.; Arap, W. Molecular addresses in blood vessels as targets for therapy. Curr. Opin. Chem. Biol. 2001, 5, 308-313. (66) Rosenwald, A.; Wright, G.; Chan, W. C.; Connors, J. M.; Campo, E.; Fisher, R. I.; Gascoyne, R. D.; Muller-Hermelink, H. K.; Smeland, E. B.; Giltnane, J. M.; Hurt, E. M.; Zhao, H.; Averett, L.; Yang, L.; Wilson, W. H.; Jaffe, E. S.; Simon, R.; Klausner, R. D.; Powell, J.; Duffey, P. L.; Longo, D. L.; Greiner, T. C.; Weisenburger, D. D.; Sanger, W. G.; Dave, B. J.; Lynch, J. C.; Vose, J.; Armitage, J. O.; Montserrat, E.; Lopez-Guillermo, A.; Grogan, T. M.; Miller, T. P.; LeBlanc, M.; Ott, G.; Kvaloy, S.; Delabie, J.; Holte, H.; Krajci, P.; Stokke, T.; Staudt, L. M. The use of molecular profiling to predict survival after chemotherapy for diffuse large-B-cell lymphoma. N. Engl. J. Med. 2002, 346, 1937-1947. (67) Liotta, L. A.; Kohn, E. C.; Petricoin, E. F. Clinical proteomics: personalized molecular medicine. JAMA 2001, 286, 2211-2214. (68) Petricoin, E. F.; Zoon, K. C.; Kohn, E. C.; Barrett, J. C.; Liotta, L. A. Clinical Proteomics: Translating Benchside Promise into Bedside Reality. Nature Rev. Drug Discuss. 2002, 1, 683-695. (69) Bray, D. Molecular Networks: The Top-Down View. Science 2003, 301, 1864-1865. (70) Ponder, B. A. Cancer genetics. Nature 2001, 411, 337-341. (71) Evan, G. I.; Vousden, K. H. Proliferation, cell cycle and apoptosis in cancer. Nature Insight. Cancer 2001, 411, 342-348. (72) Kaptain, S.; Tan, L. K.; Chen, B. Her-2/neu and breast cancer. Diag. Mol. Pathol. 2001, 10, 139-152. (73) Leyland-Jones, B. Trastuzumab: hopes and realities. Lancet Oncol. 2002, 3, 137-144. (74) Sebolt-Leopold, J. S. Development of anticancer drugs targeting the MAP kinase pathway. Oncogene 2000, 19, 6594-6599. (75) Santen, R. J.; Song, R. X.; McPherson, R.; Kumar, R.; Adam, L.; Jeng, M. H.; Yue, W. The role of mitogen-activated protein (MAP)

(76)

(77)

(78)

(79)

(80) (81)

(82)

(83)

kinase in breast cancer. J. Steroid Biochem. Mol. Biol. 2002, 80, 239-256. Thiesing, J. T.; Ohno-Jones, S.; Kolibaba, K. S.; Druker, B. J. Efficacy of STI571, an abl tyrosine kinase inhibitor, in conjunction with other antileukemic agents against bcr-abl-positive cells. Blood 2000, 96, 3195-3199. Druker, B. J.; Talpaz, M.; Resta, D. J.; Peng, B.; Buchdunger, E.; Ford, J. M.; Lydon, N. B.; Kantarjian, H.; Capdeville, R.; OhnoJones, S.; Sawyers, C. L. Efficacy and safety of a specific inhibitor of the BCR-ABL tyrosine kinase in chronic myeloid leukemia. New Engl. J. Med. 2001, 344, 1031-1037. Vlahos, C. J.; Stancato, L. F. Inhibitors of cellular signaling targetssdesigns and limitations. Submitted to Platelets and Megakaryoctyes: Methods and Protocols In: Gibbons, J. M., Mahaut-Smith, M. P., Eds.; Humana Press: Totowa, NJ, 2002. Traxler, P.; Bold, G.; Buchdunger, E.; Caravatti, G.; Furet, P.; Manley, P.; O’Reilly, T.; Wood, J.; Zimmermann, J. Tyrosine kinase inhibitors: from rational design to clinical trials. Med. Res. Rev. 2001, 21, 499-512. Zwick, E.; Bange, J.; Ullrich, A. Receptor tyrosine kinases as targets for anticancer drugs. Trends Mol. Med. 2001, 8, 17-23. Normanno, N.; Campiglio, M.; De, L. A.; Somenzi, G.; Maiello, M.; Ciardiello, F.; Gianni, L.; Salomon, D. S.; Menard, S. Cooperative inhibitory effect of ZD1839 (Iressa) in combination with trastuzumab (Herceptin) on human breast cancer cell growth. Ann. Oncol. 2002, 13, 65-72. Moasser, M. M.; Basso, A.; Averbuch, S. D.; Rosen, N. The tyrosine kinase inhibitor ZD1839 (“Iressa”) inhibits HER2-driven signaling and suppresses the growth of HER2-overexpressing tumor cells. Cancer Res. 2001, Oct 1 61(19), 7184-8. Cuello, M.; Ettenberg, S. A.; Clark, A. S.; Keane, M. M.; Posner, R. H.; Nau, M. M.; Dennis, P. A.; Lipkowitz, S. Down-regulation of the erbB-2 receptor by trastuzumab (herceptin) enhances tumor necrosis factor-related apoptosis-inducing ligand-mediated apoptosis in breast and ovarian cancer cell lines that overexpress erbB-2. Cancer Res. 2001, 61, 4892-4900.

PR049972M

Journal of Proteome Research • Vol. 3, No. 2, 2004 217