Proteomic Contributions to Personalized Cancer Care

Proteomic Contributions to PersonalizedCancer Care*John M. Koomen‡, Eric B. Haura, Gerold Bepler, Rebecca Sutphen,Elizabeth R. Remily-Wood, Kaaron Benson, Mohamad Hussein, Lori A. Hazlehurst,Timothy J. Yeatman, Lynne T. Hildreth, Thomas A. Sellers, Paul B. Jacobsen,David A. Fenstermacher, and William S. Dalton

Cancer impacts each patient and family differently. Ourcurrent understanding of the disease is primarily limitedto clinical hallmarks of cancer, but many specific molec-ular mechanisms remain elusive. Genetic markers can beused to determine predisposition to tumor development,but molecularly targeted treatment strategies that im-prove patient prognosis are not widely available for mostcancers. Individualized care plans, also described as per-sonalized medicine, still must be developed by under-standing and implementing basic science research intoclinical treatment. Proteomics holds great promise in con-tributing to the prevention and cure of cancer because itprovides unique tools for discovery of biomarkers andtherapeutic targets. As such, proteomics can help trans-late basic science discoveries into the clinical practice ofpersonalized medicine. Here we describe how biologicalmass spectrometry and proteome analysis interact withother major patient care and research initiatives and pres-ent vignettes illustrating efforts in discovery of diagnosticbiomarkers for ovarian cancer, development of treatmentstrategies in lung cancer, and monitoring prognosis andrelapse in multiple myeloma patients. Molecular & Cel-lular Proteomics 7:1780–1794, 2008.

The discovery of the causative genetic underpinnings ofcancer has been a focus of biomedical research for decades.The multigenic nature of cancer has hindered progress inunderstanding the underlying mechanisms that lead to a spe-cific disease phenotype. Recent advances in high throughputtechnologies, which evaluate tens of thousands of genes orproteins in a single experiment, are providing new methodsfor identifying biochemical determinants of the disease proc-ess. To facilitate these technologies, the correlation of spe-cific phenotypes to individual genotypes is key to leveragingthe use of model organisms and patient samples in cancerresearch. Integration of these data allows cancer researchersto ask complex questions about the mechanism of specificdisease manifestations and to retrieve data sets containing

disparate data that can be further analyzed using statisticalmethods to reveal new insights that should be furtherinvestigated.

With the comprehensive cataloging of human genes andlinks between gene function and disease, the future of med-icine looks toward mechanistic personalized medicine ap-proaches to cure diseases such as cancer. Using arrays thatcan profile gene expression, many groups have been able todefine gene expression signatures related to diagnosis (can-cer versus benign, subtype of leukemia, etc.), prognosis (like-lihood of cure), and prediction (probability of response totherapy). Although most of these approaches remain in theresearch domain, some have been thrust into the mainstreamof standard clinical practice, e.g. Oncotype DX� for predictionof breast cancer recurrence. Proteomics will be next in line todeliver new tools to help patients with cancer live longer andhave a better quality of life.

Proteomics is an emerging field that can make unique con-tributions to the prevention and cure of cancer. From strengthin protein sequence analysis to broad scale cataloging ofproteins and post-translational modifications, a wide varietyof proteomics tools are available to effect changes in patientcare. Proteomics has the advantage over genomics-basedassays because of direct examination of the molecular ma-chinery of cell physiology, including protein expression,sequence variations and isoforms, post-translational modifi-cation, and protein-protein complexes. However, certain dis-advantages also exist, including (i) stringent requirements forsample collection, preparation, and analysis, (ii) lack of am-plification procedures similar to PCR that can allow assaydevelopment using limited biological starting material, (iii) re-quirements for purification strategies to enrich samples forintended work (e.g. phosphoprotein analysis), and (iv) costsnecessary for staffing and equipping a shared resource orclinical laboratory able to perform the required assays. None-theless proteomics techniques should be implemented withbasic clinical medicine along with DNA- and/or mRNA-basedprofiling strategies to enhance cancer screening, diagnosis,treatment, and follow-up.

An overview of the potential of these cutting edge technol-ogies in the development of personalized medicine has re-cently been presented by Dalton and Friend (1). Here we build

From the H. Lee Moffitt Cancer Center and Research Institute,University of South Florida, Tampa, Florida 33612

Received, April 16, 2008, and in revised form, July 23, 2008Published, MCP Papers in Press, July 29, 2008, DOI 10.1074/

mcp.R800002-MCP200

Review

© 2008 by The American Society for Biochemistry and Molecular Biology, Inc.1780 Molecular & Cellular Proteomics 7.10This paper is available on line at http://www.mcponline.org

on that foundation and illustrate roles for proteomics in theinteraction between research and clinical practice with spe-cific vignettes. To visualize how proteomics may contributeto the development of personalized medicine, researchersmust have an understanding of the patient’s journey fromcancer diagnosis through treatment as shown in Fig. 1A.Cancer may be detected through routine check-ups, self-exams, or following the presentation of specific symptoms.Any or all of these factors may contribute to the diagnosis ofthe incoming patient. At this point, staging and molecularprofiling will also be performed using samples obtained bytumor biopsy as well as blood and/or urine collection. Thedevelopment of personalized cancer care has several goalsthat impact current and future patients: (i) identify needs ofthe individual patient, (ii) identify biomarkers to predictneeds and risks, (iii) develop and implement methods forminimally invasive patient sampling, (iv) match the righttreatment to each patient, (v) improve the performance ofclinical trials through molecular profiling, and (vi) raise thestandard of care by partnering with other hospitals andclinical care centers.

Using the diagnosis, staging, and molecular profiles, a phy-sician can assess the patient’s prognosis and predict poten-tially effective therapies. The patient should be directed totreatment regimens based on drugs with the proper mecha-nism of action. The outcome of this step is directed treatment,which optimizes the patient’s survival chances and quality oflife. Cancer survivors must be monitored for relapse or recur-rence as well as the development of new cancers; frequentlythey will (re-)enter screening programs. As a consequence ofa cancer diagnosis, members of the patient’s family and car-egivers may choose to enroll in a screening program as well.The discovery, development, and implementation of biomar-kers for ongoing monitoring are also critical to clinicalpractice.

Communication of the challenges in treatment enables re-searchers to contribute to clinical practice (Fig. 1B). The in-teraction between clinicians and researchers must be verystrong for iterative examination of clinical practice and thedevelopment of personalized medicine. Here specific casestudies illustrate potential roles for proteomics in improvingpatient care. Targeted and broad scale proteomics experi-

FIG. 1. A model for the developmentof personalized cancer care. Each pa-tient will follow a similar journey throughdiagnosis, treatment, and ongoing mon-itoring (A). The current steps in the proc-ess (in bold) are shown with their res-pective improvements expected fromongoing research. Surviving patients,their families, and caregivers often enterscreening programs. Patients who sufferrelapse or recurrence will begin theprocess again. The standard of care maychange between initial onset and re-lapse, providing additional tools for per-sonalized medicine. The interface be-tween physicians and research enablesthe continuous assessment and im-provement of clinical practice (B). By un-derstanding and evaluating challengesat each step in treatment, researcherscan suggest or provide solutions, devel-oping and implementing molecularlydriven patient care. The Roman numer-als indicate the vignettes used to illus-trate the impact that proteomics re-search can have on clinical care.

Proteomics Contributions to Personalized Cancer Care

Molecular & Cellular Proteomics 7.10 1781

ments have been implemented for the discovery of diagnosticbiomarkers in ovarian cancer (Vignette I). Phosphoproteomicscontributes to preclinical models for directed tyrosine kinaseinhibitor treatment of lung cancer (Vignette II). Detection ofdisease progression in multiple myeloma with quantitativemass spectrometry illustrates aspects of ongoing patient as-sessment (Vignette III). Finally we discuss institutional infra-structure and an example of successful implementation mo-lecular biomarkers into personalized cancer care.

VIGNETTE I: PROTEIN BIOMARKER DISCOVERY: EARLY DETECTIONOF OVARIAN CANCER

Most patients with ovarian cancer have widespread meta-static disease at initial diagnosis largely because of the inabil-ity to detect ovarian cancer at an early stage (2–5). There iscurrently no proven, effective method for early detection ofovarian cancer through biomarkers, imaging, or other means(6–11). The most commonly used biomarker for ovarian can-cer, CA125 (12), is elevated in only about 50% of stage Iovarian cancer cases (8). Beyond the lack of effective detec-tion, there is no accurate method for diagnosis of ovariancancer short of surgery; even among symptomatic women,tissue evaluation by a pathologist is the only reliable way todistinguish between women with benign and malignant dis-ease. Because of these limitations, ovarian cancer is detectedat later stages where patients have very poor prognosis andfew treatment options. Early detection significantly improvespatient outcomes.

The need for determining additional biomarkers for earlydetection of ovarian cancer that complement CA125 has beenreviewed in great detail from many perspectives, recentlyincluding proteomics. Discovery of appropriate biomarkerswould enable population screening and personalized care. Anextensive list of candidates has been prepared by Williams et al.(13) that includes proteins, glycans, lipids, and metabolites.However, Jacobs and Menon (14) describe the difficulties inher-ent in screening for ovarian cancer; in particular, the requiredspecificity would need to be essentially 100% to produce abiomarker with sufficient positive predictive value. None of theexisting candidates have been implemented in patient carestrategies, so discovery efforts continue. Although these re-views shed light on the challenges and prior candidate biomar-kers, here we illustrate how early detection of disease has beena proving ground for proteomics strategies and discuss thesuccesses of these cancer biomarker discovery efforts.

Because of the difficult nature of this task and the desire tohave unbiased approaches, researchers have applied the en-tire menu of proteomics tools to finding novel candidate bio-markers. Often these experiments are used as a provingground for analytical technology, and the rigorous require-ments for sample collection, processing, and analysis aredetermined retrospectively for improvement in the next roundof sample analysis. The ongoing development and complexityof proteomics as well as calls to standardize experiments

across institutions have raised many dilemmas for investiga-tors. The challenge in broad scale plasma proteome analysisis reflected in the fact that targeting methods rarely have thedepth to detect clinically relevant molecules released from atumor, and proteome cataloging experiments are not practicalfor case-control studies of sufficient population to detectstatistically significant differences. Furthermore each pro-teomics experiment will have specific strengths and weak-nesses because of the method of protein or peptide selectionand the type of visualization or detection chosen (15–17).

Mass spectrometry profiling and intact protein separations,including two-dimensional gel electrophoresis (2DE)1 andmultidimensional liquid chromatography, have been used totarget differences between cancer patients and controls; pro-teomes from healthy controls and cancer patients have beencataloged using LC-MS/MS shotgun sequencing. Each tech-nique will be reviewed to illustrate its strengths and the dataresources that have been produced. Although pattern analy-sis is integral to the detection and targeting of candidatebiomarkers, approaches that rely on fingerprinting alone,without identified target molecules, will be omitted becausethe addition of novel diagnostic molecules will bring the mostvalue to patient care.

Differential Display Techniques: Mass Spectrometry Profil-ing—MS profiling has been shown to be an effective methodfor detecting differences in plasma of cancer patients andcontrols. This method is attractive because the approach canbe rapid, and parallel processing enables the high throughputrequired for clinical sample analysis. Implementation appearsdeceptively simple. Despite controversy around the initial re-port using selection by surface retentate chromatography andmass analysis with MALDI MS to fingerprint ovarian cancerpatients, controls, and patients with benign disease (18), MSprofiling approaches have substantially improved even in lightof the limited number of proteins or peptides that can bedetected (19). After initial investigations of chemical fraction-ation methods using reverse phase, ion exchange, and IMAC,most of the components of the low molecular weight serum/plasma proteome were found to be intact highly abundantproteins or proteolytic fragments of plasma proteins. Thismethod of selection and detection is limited in sensitivity,peak capacity, and dynamic range; therefore, it is unlikely todetect components at levels below 1 �M unless it is combinedwith immunoprecipitation (20).

1 The abbreviations used are: 2DE, two-dimensional gel electro-phoresis; MM, multiple myeloma; MGUS, monoclonal gammopathyof undetermined significance; MRM, multiple reaction monitoring;EIC, extracted ion chromatogram; TKI, tyrosine kinase inhibitor; ITIH4,inter-�-trypsin inhibitor heavy chain H4; 2D, two-dimensional; SH2,SRC homology 2; GAP, GTPase-activating protein; STAT, signaltransducers and activators of transcription; EGFR, epidermal growthfactor receptor; ERK, extracellular signal-regulated kinase; SPEP,serum protein electrophoresis; IFE, immunofixation electrophoresis;HER2, human epidermal growth factor receptor 2.


1782 Molecular & Cellular Proteomics 7.10

Nevertheless MS profiling has detected differences be-tween ovarian cancer patients and controls, including the �

subunit of haptoglobin (21) and a panel consisting of apoli-poprotein A-I, transthyretin, and inter-�-trypsin inhibitor heavychain H4 (ITIH4) (22). Subsequent work using capture byimmobilized antibodies shows variations in ITIH4 processingin several types of cancer when compared with controls (23).Further investigation of transthyretin has revealed differencesin redox modifications including cysteinylation and glutathio-nylation (24). Clearly the extensive characterization of tar-geted molecules produces the most value because specificmolecular markers can be determined. In addition, MS profil-ing highlighted the role of protease activity in serum andplasma that may also be used to distinguish cancer patientsfrom controls (25, 26).

Differential Display Techniques: 2DE and MultidimensionalLC Protein Separations—Higher molecular weight proteinshave been analyzed by 2DE and multidimensional liquidphase separations. These approaches enable the assessmentof changes to intact molecules including proteolytic process-ing and other post-translational modifications like glycosyla-tion or phosphorylation. After protein identification, thesemodifications can be investigated further in the hope of de-veloping an assay for detecting the specific differentially ex-pressed isoforms.

In fluorescence DIGE, each sample is labeled with a differ-ent fluorescent probe, combined, and separated by isoelec-tric focusing and SDS-PAGE. Differences in protein expres-sion can be distinguished as spots that are either red or green;complete overlap is represented by yellow (Fig. 2A). Thisapproach has been popular because of the ease of interpret-ing the fluorescence images, the high number of protein spots(typically � 1,000), and the fact that the cancer and controlsamples are processed together. In addition, post-transla-tional modifications can be targeted directly in 2D gel ap-proaches using antibodies or specific fluorescent stains thatrecognize phospho- and glycoproteins.

In ovarian cancer biomarker discovery, 2DE techniqueshave been used to examine the isoforms of abundant serumproteins, indicating differences in phosphorylated fibrinogen �

(27) as well as haptoglobin and transferrin (28). Another studypresented several potential protein biomarkers, including com-plement components, serum glycoproteins, serum protease in-hibitors, transferrin, and afamin (29). The last of these markerswas further verified by ELISAs and compared with C-reactiveprotein and CA125. DIGE or targeted staining techniques in 2Dgel analysis can provide information about post-translationalmodifications to abundant proteins in the plasma.

Multidimensional liquid phase separations complement2DE. Proteome fractionation approaches have evolved to

FIG. 2. Strategies for ovarian cancer(OvCa) diagnostic biomarker discov-ery and verification. Differential displaytechniques visualizing intact proteins areattractive because candidate biomark-ers can be manually selected and statis-tically verified; examples include 2D gel(A), 2D LC (B), and single sample massspectrometry profiling. However, greaternumbers of candidate biomarkers canbe identified by LC-MS/MS shotgun se-quencing experiments. In addition topeptide catalogs, the spectral counts (C)and the average peak areas and stand-ard deviations from extracted ion chro-matograms (D) can be used to qualita-tively and quantitatively compare cancerpatients with controls as shown here forhaptoglobin. Regardless of the targetingstrategy, quantitative mass spectrome-try can be used to narrow down the listof candidates for validation (E).



include immunodepletion of the most abundant plasma pro-teins (e.g. top six or top 12 removal) followed by ion ex-change, liquid phase isoelectric focusing or chromatofocus-ing, and reverse phase separations of the lower abundancecomponents. The intensity of the proteins in the final sep-aration, whether reverse phase chromatography or SDS-PAGE, is compared with select targets for protein identifi-cation as shown in Fig. 2B. These data were acquired usingthe commercialized version of 2D protein LC (PF2D, BeckmanCoulter). The targeted fractions are recovered, digested withtrypsin, and submitted for protein identification using LC-MS/MS peptide sequencing. Current results implicate abun-dant plasma proteins as candidate markers for ovarian can-cer. Even in mouse models with enormous tumor burden, themost prevalent differences correspond to host response orabundant plasma proteins (30).

Protein Catalogs Created by Shotgun Sequencing as Re-sources for Biomarker Discovery—The role of LC-MS/MSshotgun sequencing in the proteome analysis of biofluids hascreated resources that can be exploited for biomarker discov-ery. Through a series of analytical improvements, the humanplasma proteome has been extensively cataloged by Smithand co-workers (31–33). In addition, the Human ProteomeOrganisation (HUPO) plasma proteome project has created areference for more than 3,000 proteins identified in plasma(34–36), including the corresponding gene ontology terms(37). Extensive sequencing efforts have now identified morethan 1,500 proteins from the human urinary proteome (38)using healthy samples. In addition, ascites fluid from ovariancancer patients has been extensively analyzed in a recentpublication by Gortzak-Uzan et al. (39). Each of these proteincatalogs could be scanned to reveal candidate biomarkersbased on disease etiology or organ site.

In addition to creating protein catalogs, (semi-)quantitativemeasurements can be derived from LC-MS/MS analysis ofgroups of patients and controls. The peptide counting statis-tics, which describe the number of peptides or tandem massspectra assigned to sequences from a given protein, can beused to estimate the relative amount of a protein in a complexmixture as shown for haptoglobin in Fig. 2C. The intensity ofthe intact peptide in the mass spectra can also be used toquantify the relative amount of protein in each sample; thepeak areas are calculated for each individual peptide usingextracted ion chromatograms (EICs) as shown for two pep-tides from haptoglobin in Fig. 2D. Although the primary goal ofLC-MS/MS is the catalog of proteins generated by peptidesequence assignments, EIC analysis can provide excellentdata on the relative expression level of proteins in cancerpatients’ and controls’ plasma. Furthermore the protein andits representative peptides need only be identified with highconfidence in one of the samples. With reproducible chroma-tography and accurate mass measurement, the peaks in theother samples can be correlated prior to EIC analysis. Thisapproach has been termed “label-free” proteomics, and it has

been widely and effectively used to characterize biologicaland clinical samples (40).

Narrowing the Field of Candidate Biomarkers with Quanti-tative Mass Spectrometry—A strong case can be made thatextensive verification efforts using quantitative mass spec-trometry should be the next step for biomarker developmentfor ovarian cancer. Using molecules described above in cu-rated reviews (13) or from tissue proteome analyses (41–44)investigators could apply absolute quantification to evaluatemany candidates in the same sample in a single analyticalexperiment. Peptides detected during protein identificationexperiments can be immediately used for quantitative massspectrometry analysis. Multiple reaction monitoring is typi-cally used to specifically quantify individual peptides, whichrepresent their proteins of origin (an additional description isincluded in Vignette III). Even in complex matrices like plasma,individual peptides can be monitored effectively (45–47). Mul-tiplexing strategies have also proven to be effective; Andersonand Hunter (48) used LC-multiple reaction monitoring (MRM)to develop a quantitative assay for 53 plasma proteins, illus-trating the breadth of targets that could be accessed in asingle analysis. The quantities of the proteins can be plottedby sample group, illustrating the potential utility in separat-ing cancer patients from controls, as shown in Fig. 2E.Overlapping distributions (left) will not make effective can-didates; proteins expressed at higher levels in most cancerpatients (right) can be further validated by quantitative massspectrometry or immunoassays using larger sample groups.Many of the current candidate biomarkers for detection ofovarian cancer could be better defined or ruled out usingquantitative assays, including MRM or other high resolutionLC-MS techniques.

VIGNETTE II: PRECLINICAL MODELING OF TREATMENT STRATEGIES:EXAMINATION OF ONCOGENIC TYROSINE KINASE SIGNALING AND

TYROSINE KINASE INHIBITION USING PHOSPHOPROTEOMICS

Following diagnosis, each patient is placed on a particulartreatment regimen. At present, few if any of the broadly de-ployed strategies are molecularly driven. Proteomics can beused for preclinical modeling and probing archived tissuesections for biomarkers of response or resistance. Theseinvestigations begin by matching the appropriate proteomicstools to the clinical problem and relevant biological pathways.The application of phosphoproteomics holds great promisefor understanding oncogenic signaling pathways and devel-oping biomarkers that could be predictive for patient outcomeon specific drug regimens. The focus here is to describe howproteomics technology can be applied to the study of tyrosinekinase signaling pathways and tyrosine kinase inhibitors(TKIs) in cancer. The important points are: (i) signaling path-ways are assembled in distinct modules, (ii) many of thetargets within these modules have inhibitors moving towardclinical use, (iii) assays that predict function/activity/depend-ence of these modules may allow for personalized therapy, (iv)



the existence of interchangeable modules produces complex-ity in signaling pathways, and (v) existence of redundant mod-ules and/or complex networks suggests the need for combi-natorial strategies for future clinical trials. Proteomics canhave a major impact on identifying these modules, developingpharmacodynamic assays, and unraveling the complexity ofsignaling networks.

Targeting Oncogenic Kinase Signaling Pathways—Normalcell physiology is controlled by proteins within the cell that acttogether similar to an electric circuit to ensure normal cellbehavior. Cancer is a disease where these circuits are dys-regulated or rearranged in such a way that the output drivesthe cell toward excessive growth and spread to unintendedareas of the body. Knowledge of these individual moleculesand cohesive signaling modules can help identify proteinsinvolved in driving a particular cancer cell and suggest com-binatorial therapeutic strategies. Critical components of thesecircuits are signaling proteins called kinases that act as relaysand regulate the activity of other important genes and pro-teins. Protein kinase signaling pathways regulate the “hall-marks of cancer” including cell growth, survival, invasion/metastasis, and angiogenesis (49). Not surprisingly, it hasbeen known for quite some time that aberrant kinase signalingcan lead to tumorigenesis. Notable examples of tyrosine ki-nases driving cancer are viral SRC and the breakpoint clusterregion protein and Abelson murine leukemia viral oncogenehomolog 1 (BCR-ABL) fusion protein that displays constitutivekinase activity in chronic myelogenous leukemia (50).

Considerable enthusiasm continues to focus on targetingaberrant kinase pathways in lung cancer. Both tyrosine andserine/threonine kinases are under investigation as are thepathways they regulate. Sequencing of the human genomeidentified nearly 100 tyrosine kinase proteins, some of whichare known to be involved in the pathogenesis of cancer aswell as other tyrosine kinases with potential (as yet undefined)roles. Tyrosine kinases can either span the cellular membraneand become activated by extracellular ligands (receptor tyro-sine kinase) or exist as intracellular proteins activated byintracellular events (non-receptor tyrosine kinases). The cata-lytic core subunit of tyrosine kinases recruits ATP and phos-phorylates a tyrosine residue on substrate (downstream) pro-teins. In some cases, the tyrosine kinase autophosphorylatesitself on specific sites, leading to enhanced function as well asproducing a potential biomarker for activated kinase. For ex-ample, SRC family members can both phosphorylate down-stream substrates as well as autophosphorylate itself on ty-rosine 419. Because this event leads to enhanced catalyticactivity, the degree of autophosphorylation serves as a bio-marker of SRC activity in tumor cells. Phosphorylated tyrosineresidues on substrate proteins change cellular physiology bymodifying protein functions, including enzyme activity, sub-cellular localization, and/or protein-protein interactions. Im-portant for relaying signaling, phosphorylated tyrosine sitescan act as docking sites for proteins containing SRC homol-

ogy 2 (SH2) domains (51). The human genome encodes �110distinct SH2 domain-containing proteins, and although thesedomains are generally conserved they still retain enough var-iability to lead to specificity in signaling (52). SH2-containingproteins have diverse functions including adaptors (Grb2),scaffolds (Shc), kinases (SRC), phosphatases (Shp2), Rassignaling (RasGAP), transcription (STAT), ubiquitination (Cbl),cytoskeletal function (Tensin), and phospholipid second mes-senger signaling (Polo-like kinase). Signaling originating fromtyrosine kinases pass through individual substrate proteinsand interactions with SH2 proteins, which are ultimately linkedto downstream effectors. These common sets of effectorpathways include Ras/Raf/mitogen-activated protein kinase/extracellular signal-regulated kinase kinase (MEK)/ERK sig-naling modules, STAT signaling modules, phosphatidylinosi-tol 3-kinase/Akt/mammalian target of rapamycin (mTOR)signaling molecules, protein kinase C modules, and others.These effector cascades regulate downstream proteins, someof which include transcription factors (DNA-binding proteins)that modulate gene expression. As a whole, dysregulation ofthese pathways alters cellular physiology and produces ma-lignant behavior in cells.

The expression of individual tyrosine kinases, substrates ofindividual tyrosine kinases, SH2 domain proteins, compo-nents of effector cascades, and genomic alterations in lungcancer cells allow for modular signaling networks to be cre-ated that can be unique for each particular cell or tumor. Thus,tumor cells produce complex signal transduction networksthat can be attacked at different points in an attempt to eitherkill tumor cells or revert the cells to benign function. The majorhurdle is to determine which set of signaling proteins is activeand relevant for an individual’s tumor. Thus, although we havea deeper understanding of how signaling networks are cre-ated, assays to determine the active signaling proteins in apatient’s tumor require further development.

The future of cancer treatment will be based on personal-ized approaches that identify the critical molecules necessaryfor tumor growth and survival and match patients to appro-priate molecularly directed therapy. If practicing clinicians canmatch a kinase inhibitor with an individual patient survivalwould dramatically improve, and toxicity could be reduced.Perhaps the best example is the use of imatinib for BCR-ABL-dependent chronic myelogenous leukemia (53). The successof imatinib was heavily influenced by the knowledge of BCR-ABL signaling and its critical importance to leukemia cells.Similar stories include the use of imatinib for gastrointestinalstromal tumors (50), Herceptin for HER2-overexpressingbreast cancer, and gefitinib/erlotinib for EGFR mutant-drivenlung cancers. However, given the apparent low rates of kinasemutations in the human genome from cancers, obvious mu-tations in key kinases may be rare, and a large group ofpatients may have tumors driven by kinases that are notmutated (54). In addition, patients without activating muta-tions may benefit from kinase inhibitors, such as the case of



lung cancer patients treated with erlotinib (55–57). Improvedoutcomes may also appear with logical combinations of in-hibitors that target important network “hubs” that regulatetumor survival. Complex proteomics approaches enable dis-covery of biomarker panels for tumors without obviousgenomic mutations in critical tyrosine kinases.

Identifying Biomarkers That Predict Clinical Outcome follow-ing TKI Treatment—The following sections describe tools andtechniques as well as current results that define tyrosine kinasesignaling networks before and after drug treatment and assist indevelopment of personalized therapy. A flowchart that pairsproteomics experiments with biochemistry/molecular biology,animal models, and early phase clinical trials is shown in Fig. 3.Chemical proteomics can be used to identify drug targets byaffinity chromatography and subsequent protein identification.Modulation of kinase activity can be measured by the amount ofautophosphorylation and modification of specific downstreamsubstrates using phosphotyrosine selection and LC-MS/MS;quantitative mass spectrometry measurements can also be in-corporated to evaluate the magnitude of the changes.

Identifying Novel Drug-Kinase Interactions through Chemi-cal Proteomics—Because of the conserved nature of tyrosinekinase domains in tyrosine kinases, small molecule inhibitorsdesigned to inhibit one tyrosine kinase protein can often be“dirty” molecules and have effects on other tyrosine kinase

proteins. For example, imatinib was found to have inhibitoryeffects on c-Kit, and this finding was exploited for the suc-cessful treatment of gastrointestinal stromal tumors (58).Studies examining the binding of compounds to individualtyrosine kinases reveal a spectrum of specificity ranging fromcompounds that bind to few tyrosine kinases to compoundsthat bind to numerous (�20) tyrosine kinases (59). In additionto binding partner identification, these studies also have theability to derive quantitative information regarding inhibitorbinding and selectivity (60). More recent studies have exam-ined entire libraries of tyrosine kinase inhibitors to producenovel drug-protein interactions that can be exploited for futuretherapeutic benefit. Similar studies have also highlightedthe ability of chemical proteomics approaches, derivatizingthe drugs to a stationary phase for affinity chromatography, toidentify serine/threonine kinases bound to TKIs as well asnon-kinase substrates of TKIs (61, 62). Thus, in the future asmapping of an individual’s tumor tyrosine kinase profile be-comes available, it may be possible to match tumor tyrosinekinase dependence with compounds or mixtures of com-pounds that bind and inhibit the driver kinases. It is likely thatexisting compounds have inhibitory actions beyond that oftheir original design, and these could be used for individualpatients. Adverse effects of compounds could also be relatedto off-target inhibition identified through such screens.

FIG. 3. Development of treatmentstrategies using tyrosine kinase inhib-itors. Selected TKIs identified in chemi-cal screens enter early phase clinical tri-als that evaluate safety, tolerability, andpharmacokinetics. Parallel preclinicaltesting can use chemical proteomics ap-proaches to discern putative bindingpartners. Identification of non-targetpartners may be correlated with adverseeffects identified in human trials. Puta-tive binding targets for TKIs can be ex-amined for expression in the active(functional) and phosphorylated stateusing shotgun phosphoproteomics.Next the effect of TKI on the function ofspecific targets can be evaluated usingquantitative strategies after purifyingTyr(P) (pY) peptides or specific proteins.Knowledge from early phase clinical tri-als and the achievable concentrations ofthe TKI in humans can be used to deter-mine dose-response effects on targetmodulation. Effects on individual targetscan be evaluated within the broader ef-fects on signaling networks again usingquantitative proteomics or in silico meth-ods. Finally clinical assays can be devel-oped that monitor pharmacodynamicmarkers in either tumor cells or in blood.



Use of Emerging Technologies and Biomarkers to IdentifyAberrant Kinase Signaling and Predict Response to TargetedTherapies—Identifying patients that will benefit from kinaseinhibitors remains a critical problem. Some examples of pos-sible assays to predict sensitivity to kinase inhibitors includemutation analysis on genomic DNA, evaluation of gene am-plification (fluorescence in situ hybridization), immunohisto-chemistry, and gene expression analysis (63, 64). Emergingtechnologies that produce robust proteomics analysis willfurther characterize signaling pathways that can be exploitedfor therapeutic purposes and may provide additional informa-tion relevant for patient selection and/or monitoring. MS-based proteomics may be helpful to identifying tumor cellsdependent on kinases for growth and/or survival (65). How-ever, successful implementation requires enrichment becausephosphorylated tyrosine residues (Tyr(P)) represent only 0.5%of the total phosphorylated amino acids within a cell (66).Proteomics techniques have been coupled with anti-Tyr(P)antibodies to purify Tyr(P) proteins or proteolytic Tyr(P) pep-tides for LC-MS/MS analysis (Fig. 4). Phosphotyrosine pro-teomics has been used to characterize protein networks andpathways downstream of oncogenic HER2 and BCR-ABL(67–69). These methods can also be used to identify noveltyrosine phosphorylation sites and identify oncogenic pro-teins resulting from activating mutations in protein tyrosinekinases (68–71). The data can then be used in either expertliterature curation or machine learning techniques to synthe-size network models that can be further evaluated (67). Thesemethodologies can be coupled with TKIs or other compoundsto further understand their effect on protein networks. Identi-fication of critical tyrosine kinase proteins in an importantoncogenic network may also suggest “druggable” targets thatcan be entered into therapeutic discovery research.

To illustrate the utility of such an approach, a global survey ofphosphotyrosine signaling was performed in both lung cancercell lines and primary tumors (72). This analysis identified anumber of previously identified tyrosine kinases important in

lung cancer including HER family proteins, hepatocyte growthfactor receptor, vascular endothelial growth factor receptor,IGF-1R, and SRC as well as tyrosine kinases not recognized tobe important in lung cancer pathogenesis such as human hom-olog of avian virus, anaplastic lymphoma kinase, adhesion-related kinase and platelet-derived growth factor receptor. Fromthese experiments, novel causative agents, such as fusion pro-teins incorporating ALK and ROS and aberrant platelet-derivedgrowth factor receptor � activation, were identified; furthermoresensitivity to imatinib was shown in a small subset of cell linesand tumors. Finally clustering analysis suggests distinct groupsof tumors expressing active tyrosine kinases and substrate pro-teins thus offering the possibility of identifying tumor subsetsdriven by groups of tyrosine kinases and subsets of patientsappropriate for combinations of tyrosine kinase inhibitors inclinical trials. This approach suggests that proteomics analysiscan discern individual tumor wiring circuitry that can be ex-ploited for therapeutic benefit.

As a complement to cataloging phosphotyrosine proteins,SH2 profiling can be used to examine phosphotyrosine sig-naling in cancer cells. As discussed above, tyrosine phospho-rylation on proteins serves as docking sites for proteins withSH2 domains. Screening assays using nearly the entire com-plement of human SH2 domains have been used to profilephosphotyrosine signaling in cancer cells (52). This approachwas able to discern differences in SH2 profiles (and thereforephosphotyrosine signaling) in cells transformed by distinctoncoproteins. Expansion of this approach to improve feasi-bility and enable quantitative analysis could allow for widescale profiling of cancers (73).

Quantitative Approaches to Evaluate TKI Target Modula-tion—Although identification of tyrosine kinases in the activestate in an individual tumor strongly complements studiesshowing that a particular TKI can bind that target, the tworesults do not necessarily predict that the drug can down-regulate the target activation state in vivo and that loss of a thetarget function will translate into an effect on tumor physiology

FIG. 4. Phosphotyrosine proteomicsdetects modulation of signaling afterTKI treatment. Samples are processedin parallel through lysis, protein denatur-ation, proteolysis, and phosphotyrosine(pY)-containing peptide capture. Eachsample is analyzed with LC-MS/MS on ahybrid linear ion trap-Orbitrap massspectrometer. After database searchesassign sequences of interest with highconfidence, average peak areas andstandard deviations from extracted ionchromatograms are used to examine thechanges in ion signal after TKI treatment.



(i.e. halt cell proliferation and/or initiate cell death). The typicalapproach is to treat either tumor cell lines or xenografts withTKI and then examine the effects on tumor growth along withanalysis of critical downstream signaling pathways such asphosphatidylinositol 3-kinase/Akt, STATs, and mitogen-acti-vated protein kinase (MAPK)/ERK signaling. However, suchan approach is biased because it only examines known path-ways and requires reagents including specific phosphoanti-bodies. Mass spectrometry provides an approach whereby alldetectable activated kinases and substrate phosphoproteinsare curated; the changes in all of these phosphoproteins canbe examined in the same experiment. This approach coulddiscern unique mechanisms of drug action acting throughpathways not currently believed to be either drug targets oroutside of canonical signaling pathways. In addition, off-tar-get drug effects may be elucidated that (i) predict toxicity ofcompounds and (ii) identify additional drug targets involved inother disease processes. A number of quantitative massspectrometry approaches have been reported and are con-tinuing to be improved for characterizing phosphosignalingincluding stable isotope labeling by amino acids in cell culture(SILAC), isobaric tags for relative and absolute quantification(iTRAQ), ICAT, MRM, and label-free approaches such asspectral counting and peak area calculations from EICs (74–77). One approach and representative data are shown in Fig.4. Lung cancer cell lines are treated with TKI (black) or vehiclecontrols (white). Lysates are prepared; the proteins are dena-tured, reduced, alkylated, and digested with trypsin. Afterbuffer exchange, the Tyr(P) peptides are isolated by immuno-precipitation with immobilized antibodies. After elution, cata-loging experiments are performed with LC-MS/MS. In addi-tion to sequence assignments, the phosphorylation sites areidentified. Each site can be quantified by EICs and calculatingthe total peak area for that peptide. By comparison of thepeptide ion signals, the relative amounts of phosphorylation inthe vehicle control (white) and drug treatment (black) groupscan be determined as revealed in Fig. 4.

Ultimately the incorporation of these MS-based data setsinto a coherent preclinical model will impact patient care. Theflowchart shown in Fig. 3 represents one approach incorpo-rating both preclinical and clinical information to better under-stand tyrosine kinase inhibitors and their impact on patientoutcome. Direct binding targets of tyrosine kinase inhibitorscan be discovered through chemical proteomics, and theexpression (or level of modification) of these putative targetscan be examined through shotgun phosphoproteomics. Mod-ulation of these targets can be examined through quantitativemethods described above using clinically achievable concen-trations of compounds identified through early phase clinicaltrials in patients. Inhibition of targets ultimately changes thedownstream network of signaling pathways and affects hall-marks of cancer (growth, survival, angiogenesis, etc.). Thesestudies also have the potential to lead to novel assays thatcan feed back to the clinic, such as pharmacodynamic assays

in tumor tissue or blood to examine target modulation in vivo.However, major challenges still remain. First, the amount ofdata gathered by these approaches identifies a number ofphosphotyrosine sites (or other post-translational modifica-tions) that have unclear biological significance. For some pro-teins such as EGFR and SRC, a large body of published workhas defined the functional significance of each Tyr(P) site. Formany other proteins, this is not the case. Thus, more basicbiochemical experiments are still necessary and important todefine the functional significance of certain post-translationalmodifications. Second, the interactions between differentphosphoproteins in the signaling networks are poorly defined.Cancer circuitry likely has tremendous plasticity that allowsfor rewiring depending on changes in expression patterns,localization, environmental cues, and post-translational mod-ifications. Even networks downstream of well studied tyrosinekinases such as EGFR and SRC can be complicated. Third,use of MS-based assays in patient materials is difficult giventhe limited amount of tumor obtained from core biopsies andthe large amounts of starting material required for immunoaf-finity purification of tyrosine phosphorylated peptides. Tumorcollection procedures are also critical to maintain post-trans-lational modifications, such as phosphorylation; furthermoretumor heterogeneity may complicate interpretation of results.Methods for assessing each cell in a population and samplingsmall numbers of cells are just now becoming available, in-cluding phosphospecific flow cytometry (78) and capillaryisoelectric focusing coupled to antibody-based detection (79).Thus, translation of MS-based assays to other user-friendlyassays with robust quantitative characteristics will be impor-tant to develop and validate for use in the clinical setting.

VIGNETTE III: PATIENT MONITORING WITH QUANTITATIVE MASSSPECTROMETRY: ASSESSMENT OF MYELOMA PROGRESSION

AND RELAPSE

Following treatment and partial or complete remission, pa-tients will be reassessed at regular intervals to check fordisease relapse or recurrence. Clinical assays based on quan-titative mass spectrometry can also play a role in ongoingpatient management as illustrated with monitoring multiplemyeloma (MM), which is a cancer of the plasma cell thatdevelops primarily in the elderly population. The progressionof the tumor is well understood, and it can be diagnosed bythe presence of multiple myeloma cells in the bone marrowand detected by the amount of antibody secretion from theclonal population of plasma cells. A premalignant conditionknown as monoclonal gammopathy of undetermined signifi-cance (MGUS) develops at certain rates in the United Statespopulation: 3% at age 50, 5% at age 70, and 7% by age 85;�1% of MGUS patients progress to multiple myeloma on anannual basis (80). The molecular causes for progression fromMGUS to MM are unknown. After the onset of the cancer,multiple myeloma patients suffer from several symptoms, in-cluding calcium dysregulation, renal failure, anemia, and bone



lesions. A diagnosis of multiple myeloma is established usingblood and urine tests. For advanced stage patients, completeskeletal surveys are also used to examine the damage causedby multiple myeloma in the bone marrow. Staging with serumcalcium, creatinine, hemoglobin, and most importantly theconcentration of the “monoclonal serum protein” was estab-lished in 1975 by Durie and Salmon (81). The InternationalStaging System determined in 2005 uses those markers aswell as serum albumin and �2-microglobulin (82). The survivalstatistics indicate the importance of early detection andproper staging and show the devastating impact of multiplemyeloma. Stage I patients have median survival times of 62months, stage II patients have median survival times of 45months, and stage III patient median survival is reduced to29 months (82). Novel therapeutic strategies include protea-some inhibition with agents like bortezomib (83, 84), whichcan be used as a targeted therapy; treatment is more effectivefor patients with myelomas that secrete high levels of mono-clonal antibodies (85). Patient monitoring strategies presentsignificant challenges particularly in the detection of MGUSpatients most likely to develop multiple myeloma and ongoingassessment of relapse or recurrence in previously treatedmultiple myeloma patients. Many MM patients who have un-dergone treatment are repetitively checked at 2–4-week in-tervals, leading to high numbers of clinic visits and collectionof large volumes of blood.

In the following paragraphs, methods for patient samplingand detection of the monoclonal serum protein are presentedfrom a process chemistry standpoint. Process chemists useextensive background knowledge of synthesis, analysis, andengineering to redesign industrial assembly lines or improveindividual steps in manufacturing. In this case, the currentclinical methods for analyzing serum from multiple myelomapatients are reviewed; a quantitative mass spectrometry as-say for monoclonal proteins is developed and assessed for itsvalue in clinical implementation.

Because each plasma cell secretes a unique antibody, thereplication of the tumor cell and the progression of the dis-ease can be monitored by measuring the serum concentrationof the monoclonal antibody it produces. Initial qualitativemeasurements are made using serum protein electrophoresis(SPEP) and dye visualization (see Fig. 5A). Separation of theserum proteins is achieved, isolating albumin from four re-gions of globulins, termed �1, �2, �, and �, described by thedifferences in their migration relative to albumin. Normallyantibodies migrate into the � region but are low in intensitycompared with albumin and present only as diffuse bands(Fig. 5A, left). The monoclonal antibodies produced in highconcentration by multiple myeloma cells can be visualized asa single narrow, discrete, dark band usually in the � region ofthe gel (Fig. 5A, right). Patients with abnormally high levels ofprotein in the � region can be diagnosed with multiple my-eloma after the type of antibody is defined as monoclonalusing immunofixation electrophoresis (IFE), which is a sepa-

ration similar to SPEP but with specific detection for eachantibody chain (Fig. 5B). Typical screens test for immunoglob-ulin G, A, and M heavy chains as well as � and � light chains.Immunoglobulin D or E myelomas are very rare; when sus-pected, lanes of the standard IFE are replaced, enabling spe-cific detection of IgD or IgE heavy chain proteins. In theexample, the patient has a tumor that produces an IgG �

monoclonal protein (Fig. 5B). The combination of these twotests establishes the relative amount and type of the antibodythat is secreted by the multiple myeloma tumor cells. Thesegel-based techniques have recently been complemented bycapillary array instruments that can analyze eight samples inparallel, greatly increasing the throughput and lowering theamount of sample preparation necessary (tubes of serum aresimply loaded into the instrument, which automatically diluteseach sample in the buffer used for capillary electrophoresis).Even using this new technology, SPEP and IFE are performedseparately.

A quantitative mass spectrometry method could replacethese methods with a single analysis; this technique is thesubject of a forthcoming manuscript.2 Example bands fromSPEP have been processed for protein identification usingLC-MS/MS. The detection of peptides from the constant re-gions of the antibody indicates the peptides that could bedeveloped for quantitative monitoring as shown in Fig. 5C forALPAPIEK detected from immunoglobulin G heavy chains.After generating peptides for monitoring each of the types ofantibodies, a comprehensive method for antibody measure-ment was made. Briefly minute volumes of serum (1–10 �l)can be processed for detection of each of the antibodychains: G, A, M, �, and �. After protein denaturation with urea,disulfide reduction, and cysteine alkylation, trypsin digestionis performed. The sample is then diluted and analyzed withLC-MRM on a triple quadrupole mass spectrometer (Fig. 5D).The instrument selectively quantifies peptides by filtering them/z of the intact species in the first quadrupole (Q1), frag-menting the molecules in the second quadrupole (Q2), andfiltering the m/z of a particular fragment in the third quadru-pole (Q3). Each of these peptide and fragment pairs is knownas a transition; the instrument measures each transition aspart of a cycle, continuously moving from one to the next. Foreach peptide, multiple transitions are monitored; the coinci-dence detection of multiple fragments from the peptide in-creases the confidence in the assignment. Each proteinshould be quantified using more than one peptide. Althoughseveral rules for peptide selection have been put forward,selection of peptide in biological or clinical context frequentlydeviates from those guidelines. Examples of quantificationwith LC-MRM are shown in Fig. 5, E and F, for the ALPAPIEKpeptide from immunoglobulin G and DSTYSLSSTLTLK fromthe � light chain. The ion signals corresponding to the y5 and

2 E. R. Remily, K. Benson, M. Hussein, L. A. Hazlehurst, and J. M.Koomen, unpublished data.



y6 ions of ALPAPIEK were detected at 22 min in Fig. 5E; y4 isalso monitored. The y8, y11, and y12 ion signals forDSTYSLSSTLTLK were detected at 28 min as shown in Fig.5F. These ion signals were obtained from the sample used forthe SPEP and IFE, illustrated with the schematic diagrams inFig. 5, A and B. This quantitative mass spectrometry tech-nique is currently being implemented in parallel with SPEPand IFE in a clinical trial.

Pending successful completion of this evaluation, the quan-titative mass spectrometry assay will be used in a broaderresearch context. Its use would be clearly advantageous inanimal models where limited amounts of blood serum can beobtained. However, because of the ease of use of SPEPcombined with IFE, the clinical impact of such a test may belimited because of the costs associated with and the training

necessary for successful mass spectrometry analysis. In thelong run, as robotics and automated sample processing be-come less expensive and more widely applied, this quantita-tive mass spectrometry assay could be deployed in the clinic.The implementation of a single quantitative test would provideadvantages over the qualitative tests currently used to followmultiple myeloma patients. The speed and parallel processingthat could be achieved with automated sample handling andMS detection would also significantly improve the throughputof patient samples. The adoption of this technique at a tertiarycancer center could enable surrounding primary care physi-cians and hospitals to send samples to a centralized facilityfor processing and analysis. Point of care patient samplingcould be performed with rapid turnaround of results to thetreating physician (�1 day) even at a centralized facility.

FIG. 5. Multiple myeloma diagnosisand prognosis with a quantitativemass spectrometry assay. Currentlyexcessive antibody production is de-tected with serum protein electrophore-sis (A). The antibodies are identified us-ing immunofixation electrophoresis (B);here the monoclonal spike is an IgGwith � light chain. After LC-MS/MS,peptides, like ALPAPIEK from IgG, canbe selected for quantitative mass spec-trometry assays (C). A schematic dia-gram of selected reaction monitoring isshown (D); a single molecule can bedetected by filtering the m/z values forpeptide and specific fragments. Usingthe same serum sample shown in B, thequantity and type of antibody are de-termined in a mass spectrometry as-say; high levels of ALPAPIEK from IgG(E) and DSTYSLSSTLTLSK (F) from �light chain are confirmed using multiplereaction monitoring.



CONCLUSIONS

Proteomics can have great impact on clinical cancer care inthe near future. To be successful, experimental design re-mains the most important step. To answer a specific question,the right methods have to be selected; the required speci-mens must be collected and stored with the proper tech-niques. Although a wide variety of tools and techniques areavailable, directed analysis gives the best chance for pro-teomics to assist in the delivery of personalized medicine. Incancer, both success, in the form of survivors, and failure, aspatients that succumb to disease, must be measured asclinical practice is iteratively improved.

Potential roles for proteomics have been presented usingdifferent clinical vignettes. Impact can be made in diverseareas including biomarker discovery, molecular prescriptionof targeted therapy regimens, and patient monitoring. Each ofthese descriptions should indicate the difficulty associatedwith clinical proteomics as well as the importance of closecollaboration between researchers and clinicians. The discov-ery of novel biomarkers is extremely challenging even in thestage of experiment design and technique selection. Phos-phoproteomics is making contributions to preclinical model-ing for molecularly driven therapy in lung cancer; the combi-nation of chemical proteomics, phosphoproteomics, and SH2interactions provided detailed mechanistic views of the pro-tein-protein interactions and signaling networks targeted bytyrosine kinase inhibitors. Finally quantitative mass spectrom-etry can change paradigms in patient sampling for monitoringbiomarkers as shown for multiple myeloma. Even with a di-agnostic and prognostic biomarker for this disease, manychallenges remain for treatment and patient care.

All of these research efforts move forward because of in-stitutional initiatives and support as well as strong communityinvolvement. The contributions of the patients and their fam-ilies should not be overlooked; many consent to tissue col-lection and data sharing as they enter cancer prevention,screening, and treatment regimens. At the H. Lee MoffittCancer Center and Research Institute, Lifetime CancerScreening and Total Cancer Care initiatives enroll healthycontrols and cancer patients, respectively. Data fromhealthy people that develop cancer and those that survivethe disease create powerful resources for research oppor-tunities in discovery, translation, and delivery of personal-ized cancer care. Methods for specimen banking and coa-lescing clinical and research data play an important role inthese institutional initiatives; similar to the evaluation ofpatient care, these institutional protocols must be iterativelyevaluated and improved.

Translating protein-based biomarkers into the clinic re-quires a close collaboration between oncologists, patholo-gists, statisticians, and clinical trials staff along with infra-structure for real time biomarker analysis. Such an approachcan be observed in a recent study using tailored chemother-

apy for the treatment of advanced lung cancer based on theexpression of two genes (86). Trial participation required adedicated biopsy of the tumor specifically for gene expressionanalysis performed by real time quantitative PCR. Predeter-mined values for gene expression were used for decisionsregarding usage of the chemotherapy drugs. The goal of thisstudy was not to compare different treatments but to demon-strate that upfront patient selection with tailored treatment willresult in improved outcome. This was found to be the casewith data indicating an unprecedented 12-month improve-ment in survival. The conclusion from this experience is thatreal time therapeutic decision making based on biomarkersfor patients with advanced non-small cell lung cancer is fea-sible and promising for improvement in patient outcome.However, this approach currently represents a boutique ther-apy because the expression analysis techniques used requirea substantial infrastructure and, therefore, may not be readilyaccessible to the vast majority of patients. Thus, the devel-opment of a more generally applicable methodology based ontechnology familiar to clinical laboratories and pathologists,such as immunohistochemistry, is desirable. To overcomethese limitations, using an automated, quantitative, immun-ofluorescence-based technique (accurate quantitative analy-sis (AQUA)), protein expression (rather than mRNA) was as-sessed and developed into a more applicable assay systemfor further development (87).

However, several issues should be addressed before ageneral recommendation for implementation of a biomarker-based therapeutic approach can be given. First, it is crucial tocorroborate the prognostic and predictive impact of thesebiomarkers in independent data sets under controlled condi-tions. Second, it is important to test the general feasibility ofa tailored treatment approach in a multicenter trial and toverify promising phase II outcome data in a large community-based trial. Thus, randomized and multisite clinical trials arenecessary to produce more convincing data suggesting theutility of such an approach.

Acknowledgments—We thank Bin Fang, Wei Guan, VictoriaIzumi, Umut Oguz, C. Eric Thomas, the Haura laboratory, TriciaHoltje, and Amy Koomen for contributions. Proteomics at Moffitt issupported by the United States Army Medical Research and Ma-teriel Command under Award DAMD17-02-2-0051 for a NationalFunctional Genomics Center, the NCI, National Institutes of Healthunder Award P30-CA076292 as a Cancer Center Support Grant,and the Moffitt Foundation.

* This work was supported, in whole or in part, by National Insti-tutes of Health Grants R01-CA106414 (to R. S. ), R01-CA123174 (toE. B. H.), and R01-CA102726 and U01-CA101222 (to G. B.) from theNCI. This work was also supported by the National FunctionalGenomics Center (to E. B. H. and J. M. K.) and American CancerSociety Grant CRTG-00-196-01-CCE (to R. S. ). The costs of publi-cation of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked “advertise-ment” in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.



‡ To whom correspondence should be addressed: SRB3 Molec-ular Oncology and Proteomics, H. Lee Moffitt Cancer Center andResearch Inst., University of South Florida, 12902 Magnolia Dr.,Tampa, FL 33612. Tel.: 813-745-8524; Fax: 813-745-3829; E-mail:[email protected].

REFERENCES

1. Dalton, W. S., and Friend, S. H. (2006) Cancer biomarkers—an invitation tothe table. Science 312, 1165–1168

2. American Cancer Society (2008) 2008 Estimated US Cancer Deaths, inCancer Statistics 2008, American Cancer Society, Inc., Atlanta, GA

3. Holschneider, C. H., and Berek, J. S. (2000) Ovarian cancer: epidemiology,biology, and prognostic factors. Semin. Surg. Oncol. 19, 3–10

4. Boente, M. P., Schilder, R., and Ozols, R. F. (1999) Gynecological cancers.Cancer Chemother. Biol. Response Modif. 18, 418–434

5. Baker, T. R., and Piver, M. S. (1994) Etiology, biology, and epidemiology ofovarian cancer. Semin. Surg. Oncol. 10, 242–248

6. Schwartz, P. E., and Taylor, K. J. (1995) Is early detection of ovarian cancerpossible? Ann. Med. 27, 519–528

7. Jacobs, I., and Bast, R. C., Jr. (1989) The CA 125 tumour-associatedantigen: a review of the literature. Hum. Reprod. 4, 1–12

8. Woolas, R. P., Xu, F. J., Jacobs, I. J., Yu, Y. H., Daly, L., Berchuck, A.,Soper, J. T., Clarke-Pearson, D. L., Oram, D. H., and Bast, R. C., Jr.(1993) Elevation of multiple serum markers in patients with stage Iovarian cancer. J. Natl. Cancer Inst. 85, 1748–1751

9. van Haaften-Day, C., Shen, Y., Xu, F., Yu, Y., Berchuck, A., Havrilesky, L. J.,de Bruijn, H. W., van der Zee, A. G., Bast, R. C., Jr., and Hacker, N. F.(2001) OVX1, macrophage-colony stimulating factor, and CA-125-II astumor markers for epithelial ovarian carcinoma: a critical appraisal. Can-cer 92, 2837–2844

10. Taylor, K. J., and Schwartz, P. E. (1994) Screening for early ovarian cancer.Radiology 192, 1–10

11. Bast, R. C., Jr., Xu, F. J., Yu, Y. H., Barnhill, S., Zhang, Z., and Mills, G. B.(1998) CA 125: the past and the future. Int. J. Biol. Markers 13, 179–187

12. Bast, R. C., Jr., Feeney, M., Lazarus, H., Nadler, L. M., Colvin, R. B., andKnapp, R. C. (1981) Reactivity of a monoclonal antibody with humanovarian carcinoma. J. Clin. Investig. 68, 1331–1337

13. Williams, T. I., Toups, K. L., Saggese, D. A., Kalli, K. R., Cliby, W. A., andMuddiman, D. C. (2007) Epithelial ovarian cancer: disease etiology, treat-ment, detection, and investigational gene, metabolite, and protein bi-omarkers. J. Proteome Res. 6, 2936–2962

14. Jacobs, I. J., and Menon, U. (2004) Progress and challenges in screeningfor early detection of ovarian cancer. Mol. Cell. Proteomics. 3, 355–366

15. Faca, V., Pitteri, S. J., Newcomb, L., Glukhova, V., Phanstiel, D., Krasnosel-sky, A., Zhang, Q., Struthers, J., Wang, H., Eng, J., Fitzgibbon, M.,McIntosh, M., and Hanash, S. (2007) Contribution of protein fractionationto depth of analysis of the serum and plasma proteomes. J. ProteomeRes. 6, 3558–3565

16. Whiteaker, J. R., Zhang, H., Eng, J. K., Fang, R., Piening, B. D., Feng, L. C.,Lorentzen, T. D., Schoenherr, R. M., Keane, J. F., Holzman, T., Fitzgib-bon, M., Lin, C., Zhang, H., Cooke, K., Liu, T., Camp, D. G., II, Anderson,L., Watts, J., Smith, R. D., McIntosh, M. W., and Paulovich, A. G. (2007)Head-to-head comparison of serum fractionation techniques. J. Pro-teome Res. 6, 828–836

17. Li, X., Gong, Y., Wang, Y., Wu, S., Cai, Y., He, P., Lu, Z., Ying, W., Zhang,Y., Jiao, L., He, H., Zhang, Z., He, F., Zhao, X., and Qian, X. (2005)Comparison of alternative analytical techniques for the characterisationof the human serum proteome in HUPO Plasma Proteome Project.Proteomics 5, 3423–3441

18. 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.,and Liotta, L. A. (2002) Use of proteomic patterns in serum to identifyovarian cancer. Lancet 359, 572–577

19. Koomen, J. M., Li, D., Xiao, L. C., Liu, T. C., Coombes, K. R., Abbruzzese,J., and Kobayashi, R. (2005) Direct tandem mass spectrometry revealslimitations in protein profiling experiments for plasma biomarker discov-ery. J. Proteome Res. 4, 972–981

20. Nedelkov, D., Kiernan, U. A., Niederkofler, E. E., Tubbs, K. A., and Nelson,R. W. (2005) Investigating diversity in human plasma proteins. Proc. Natl.Acad. Sci. U. S. A. 102, 10852–10857

21. Ye, B., Cramer, D. W., Skates, S. J., Gygi, S. P., Pratomo, V., Fu, L., Horick,

N. K., Licklider, L. J., Schorge, J. O., Berkowitz, R. S., and Mok, S. C.(2003) Haptoglobin-� subunit as potential serum biomarker in ovariancancer: identification and characterization using proteomic profiling andmass spectrometry. Clin. Cancer Res. 9, 2904–2911

22. Zhang, Z., Bast, R. C., Jr., Yu, Y., Li, J., Sokoll, L. J., Rai, A. J., Rosenzweig,J. M., Cameron, B., Wang, Y. Y., Meng, X. Y., Berchuck, A., Van Haaften-Day, C., Hacker, N. F., de Bruijn, H. W., van der Zee, A. G., Jacobs, I. J.,Fung, E. T., and Chan, D. W. (2004) Three biomarkers identified fromserum proteomic analysis for the detection of early stage ovarian cancer.Cancer Res. 64, 5882–5890

23. Song, J., Patel, M., Rosenzweig, C. N., Chan-Li, Y., Sokoll, L. J., Fung, E. T.,Choi-Miura, N. H., Goggins, M., Chan, D. W., and Zhang, Z. (2006)Quantification of fragments of human serum inter-�-trypsin inhibitorheavy chain 4 by a surface-enhanced laser desorption/ionization-basedimmunoassay. Clin. Chem. 52, 1045–1053

24. Fung, E. T., Yip, T. T., Lomas, L., Wang, Z., Yip, C., Meng, X. Y., Lin, S.,Zhang, F., Zhang, Z., Chan, D. W., and Weinberger, S. R. (2005) Classi-fication of cancer types by measuring variants of host response proteinsusing SELDI serum assays. Int. J. Cancer 115, 783–789

25. Villanueva, J., Shaffer, D. R., Philip, J., Chaparro, C. A., Erdjument-Bro-mage, H., Olshen, A. B., Fleisher, M., Lilja, H., Brogi, E., Boyd, J.,Sanchez-Carbayo, M., Holland, E. C., Cordon-Cardo, C., Scher, H. I.,and Tempst, P. (2006) Differential exoprotease activities confer tumor-specific serum peptidome patterns. J. Clin. Investig. 116, 271–284

26. Villanueva, J., Nazarian, A., Lawlor, K., Yi, S. S., Robbins, R. J., and Tempst,P. (2008) A sequence-specific exopeptidase activity test (SSEAT) for“functional” biomarker discovery. Mol. Cell. Proteomics 7, 509–518

27. Ogata, Y., Heppelmann, C. J., Charlesworth, M. C., Madden, B. J., Miller,M. N., Kalli, K. R., Cilby, W. A., Bergen, H. R., III, Saggese, D. A., andMuddiman, D. C. (2006) Elevated levels of phosphorylated fibrinogen-�-isoforms and differential expression of other post-translationally modi-fied proteins in the plasma of ovarian cancer patients. J. Proteome Res.5, 3318–3325

28. Ahmed, N., Oliva, K. T., Barker, G., Hoffmann, P., Reeve, S., Smith, I. A.,Quinn, M. A., and Rice, G. E. (2005) Proteomic tracking of serum proteinisoforms as screening biomarkers of ovarian cancer. Proteomics 5,4625–4636

29. Jackson, D., Craven, R. A., Hutson, R. C., Graze, I., Lueth, P., Tonge, R. P.,Hartley, J. L., Nickson, J. A., Rayner, S. J., Johnston, C., Dieplinger, B.,Hubalek, M., Wilkinson, N., Perren, T. J., Kehoe, S., Hall, G. D., Daxen-bichler, G., Dieplinger, H., Selby, P. J., and Banks, R. E. (2007) Proteomicprofiling identifies afamin as a potential biomarker for ovarian cancer.Clin. Cancer Res. 13, 7370–7379

30. Kuick, R., Misek, D. E., Monsma, D. J., Webb, C. P., Wang, H., Peterson,K. J., Pisano, M., Omenn, G. S., and Hanash, S. M. (2007) Discovery ofcancer biomarkers through the use of mouse models. Cancer Lett. 249,40–48

31. Shen, Y., Jacobs, J. M., Camp, D. G., II, Fang, R., Moore, R. J., Smith, R. D.,Xiao, W., Davis, R. W., and Tompkins, R. G. (2004) Ultra-high-efficiencystrong cation exchange LC/RPLC/MS/MS for high dynamic range char-acterization of the human plasma proteome. Anal. Chem. 76, 1134–1144

32. Shen, Y., Kim, J., Strittmatter, E. F., Jacobs, J. M., Camp, D. G., II, Fang, R.,Tolie, N., Moore, R. J., and Smith, R. D. (2005) Characterization of thehuman blood plasma proteome. Proteomics 5, 4034–4045

33. Adkins, J. N., Varnum, S. M., Auberry, K. J., Moore, R. J., Angell, N. H.,Smith, R. D., Springer, D. L., and Pounds, J. G. (2002) Toward a humanblood serum proteome: analysis by multidimensional separation coupledwith mass spectrometry. Mol. Cell. Proteomics 1, 947–955

34. Omenn, G. S., States, D. J., Adamski, M., Blackwell, T. W., Menon, R.,Hermjakob, H., Apweiler, R., Haab, B. B., Simpson, R. J., Eddes, J. S.,Kapp, E. A., Moritz, R. L., Chan, D. W., Rai, A. J., Admon, A., Aebersold,R., Eng, J., Hancock, W. S., Hefta, S. A., Meyer, H., Paik, Y. K., Yoo, J. S.,Ping, P., Pounds, J., Adkins, J., Qian, X., Wang, R., Wasinger, V., Wu,C. Y., Zhao, X., Zeng, R., Archakov, A., Tsugita, A., Beer, I., Pandey, A.,Pisano, M., Andrews, P., Tammen, H., Speicher, D. W., and Hanash,S. M. (2005) Overview of the HUPO Plasma Proteome Project: resultsfrom the pilot phase with 35 collaborating laboratories and multipleanalytical groups, generating a core dataset of 3020 proteins and apublicly-available database. Proteomics 5, 3226–3245

35. States, D. J., Omenn, G. S., Blackwell, T. W., Fermin, D., Eng, J., Speicher,D. W., and Hanash, S. M. (2006) Challenges in deriving high-confidence



protein identifications from data gathered by a HUPO plasma proteomecollaborative study. Nat. Biotechnol. 24, 333–338

36. Adkins, J. N., Monroe, M. E., Auberry, K. J., Shen, Y., Jacobs, J. M., Camp,D. G., II, Vitzthum, F., Rodland, K. D., Zangar, R. C., Smith, R. D., andPounds, J. G. (2005) A proteomic study of the HUPO Plasma ProteomeProject’s pilot samples using an accurate mass and time tag strategy.Proteomics 5, 3454–3466

37. Ping, P., Vondriska, T. M., Creighton, C. J., Gandhi, T. K., Yang, Z., Menon,R., Kwon, M. S., Cho, S. Y., Drwal, G., Kellmann, M., Peri, S., Suresh, S.,Gronborg, M., Molina, H., Chaerkady, R., Rekha, B., Shet, A. S., Gersz-ten, R. E., Wu, H., Raftery, M., Wasinger, V., Schulz-Knappe, P., Hanash,S. M., Paik, Y. K., Hancock, W. S., States, D. J., Omenn, G. S., andPandey, A. (2005) A functional annotation of subproteomes in humanplasma. Proteomics 5, 3506–3519

38. Adachi, J., Kumar, C., Zhang, Y., Olsen, J. V., and Mann, M. (2006) Thehuman urinary proteome contains more than 1500 proteins, including alarge proportion of membrane proteins. Genome Biol. 7, R80

39. Gortzak-Uzan, L., Ignatchenko, A., Evangelou, A. I., Agochiya, M., Brown,K. A., St Onge, P., Kireeva, I., Schmitt-Ulms, G., Brown, T. J., Murphy, J.,Rosen, B., Shaw, P., Jurisica, I., and Kislinger, T. (2008) A proteomeresource of ovarian cancer ascites: integrated proteomic and bioinfor-matic analyses to identify putative biomarkers. J. Proteome Res. 7,339–351

40. Wu, S. L., Choudhary, G., Ramstrom, M., Bergquist, J., and Hancock, W. S.(2003) Evaluation of shotgun sequencing for proteomic analysis of hu-man plasma using HPLC coupled with either ion trap or Fourier transformmass spectrometry. J. Proteome Res. 2, 383–393

41. Alaiya, A. A., Franzen, B., Fujioka, K., Moberger, B., Schedvins, K., Silfvers-vard, C., Linder, S., and Auer, G. (1997) Phenotypic analysis of ovariancarcinoma: polypeptide expression in benign, borderline and malignanttumors. Int. J. Cancer 73, 678–683

42. Alaiya, A. A., Franzen, B., Moberger, B., Silfversward, C., Linder, S., andAuer, G. (1999) Two-dimensional gel analysis of protein expression inovarian tumors shows a low degree of intratumoral heterogeneity. Elec-trophoresis 20, 1039–1046

43. Bengtsson, S., Krogh, M., Szigyarto, C. A., Uhlen, M., Schedvins, K.,Silfversward, C., Linder, S., Auer, G., Alaiya, A., and James, P. (2007)Large-scale proteomics analysis of human ovarian cancer for biomark-ers. J. Proteome Res. 6, 1440–1450

44. Wang, H., Kachman, M. T., Schwartz, D. R., Cho, K. R., and Lubman, D. M.(2004) Comprehensive proteome analysis of ovarian cancers using liquidphase separation, mass mapping and tandem mass spectrometry: astrategy for identification of candidate cancer biomarkers. Proteomics 4,2476–2495

45. Kuhn, E., Wu, J., Karl, J., Liao, H., Zolg, W., and Guild, B. (2004) Quantifi-cation of C-reactive protein in the serum of patients with rheumatoidarthritis using multiple reaction monitoring mass spectrometry and 13C-labeled peptide standards. Proteomics 4, 1175–1186

46. Hawkridge, A. M., Heublein, D. M., Bergen, H. R., III, Cataliotti, A., Burnett,J. C., Jr., and Muddiman, D. C. (2005) Quantitative mass spectral evi-dence for the absence of circulating brain natriuretic peptide (BNP-32) insevere human heart failure. Proc. Natl. Acad. Sci. U. S. A. 102,17442–17447

47. Melanson, J. E., Chisholm, K. A., and Pinto, D. M. (2006) Targeted com-parative proteomics by liquid chromatography/matrix-assisted laser de-sorption/ionization triple-quadrupole mass spectrometry. Rapid Com-mun. Mass Spectrom. 20, 904–910

48. Anderson, L., and Hunter, C. L. (2006) Quantitative mass spectrometricmultiple reaction monitoring assays for major plasma proteins. Mol. Cell.Proteomics 5, 573–588

49. Blume-Jensen, P., and Hunter, T. (2001) Oncogenic kinase signaling. Na-ture 411, 355–365

50. Krause, D. S., and Van Etten, R. A. (2005) Tyrosine kinases as targets forcancer therapy. N. Engl. J. Med. 353, 172–187

51. Pawson, T. (2004) Specificity in signal transduction: from phosphotyrosine-SH2 domain interactions to complex cellular systems. Cell 116, 191–203

52. Machida, K., Thompson, C. M., Dierck, K., Jablonowski, K., Karkkainen, S.,Liu, B., Zhang, H., Nash, P. D., Newman, D. K., Nollau, P., Pawson, T.,Renkema, G. H., Saksela, K., Schiller, M. R., Shin, D. G., and Mayer, B. J.(2007) High-throughput phosphotyrosine profiling using SH2 domains.Mol. Cell 26, 899–915

53. Druker, B. J. (2003) David A. Karnofsky Award lecture. Imatinib as aparadigm of targeted therapies. J. Clin. Oncol. 21, (suppl.) 239s–245s.

54. Davies, H., Hunter, C., Smith, R., Stephens, P., Greenman, C., Bignell, G.,Teague, J., Butler, A., Edkins, S., Stevens, C., Parker, A., O’Meara, S.,Avis, T., Barthorpe, S., Brackenbury, L., Buck, G., Clements, J., Cole, J.,Dicks, E., Edwards, K., Forbes, S., Gorton, M., Gray, K., Halliday, K.,Harrison, R., Hills, K., Hinton, J., Jones, D., Kosmidou, V., Laman, R.,Lugg, R., Menzies, A., Perry, J., Petty, R., Raine, K., Shepherd, R., Small,A., Solomon, H., Stephens, Y., Tofts, C., Varian, J., Webb, A., West, S.,Widaa, S., Yates, A., Brasseur, F., Cooper, C. S., Flanagan, A. M., Green,A., Knowles, M., Leung, S. Y., Looijenga, L. H., Malkowicz, B., Pierotti,M. A., Teh, B. T., Yuen, S. T., Lakhani, S. R., Easton, D. F., Weber, B. L.,Goldstraw, P., Nicholson, A. G., Wooster, R., Stratton, M. R., and Futreal,P. A. (2005) Somatic mutations of the protein kinase gene family inhuman lung cancer. Cancer Res. 65, 7591–7595

55. Kris, M. G., Natale, R. B., Herbst, R. S., Lynch, T. J., Jr., Prager, D., Belani,C. P., Schiller, J. H., Kelly, K., Spiridonidis, H., Sandler, A., Albain, K. S.,Cella, D., Wolf, M. K., Averbuch, S. D., Ochs, J. J., and Kay, A. C. (2003)Efficacy of gefitinib, an inhibitor of the epidermal growth factor receptortyrosine kinase, in symptomatic patients with non-small cell lung cancer:a randomized trial. J. Am. Med. Assoc. 290, 2149–2158

56. Perez-Soler, R., Chachoua, A., Hammond, L. A., Rowinsky, E. K., Huber-man, M., Karp, D., Rigas, J., Clark, G. M., Santabarbara, P., and Bonomi,P. (2004) Determinants of tumor response and survival with erlotinib inpatients with non-small-cell lung cancer. J. Clin. Oncol. 22, 3238–3247

57. Shepherd, F. A., Rodrigues Pereira, J., Ciuleanu, T., Tan, E. H., Hirsh, V.,Thongprasert, S., Campos, D., Maoleekoonpiroj, S., Smylie, M., Martins,R., van Kooten, M., Dediu, M., Findlay, B., Tu, D., Johnston, D., Bezjak,A., Clark, G., Santabarbara, P., Seymour, L., and the National CancerInstitute of Canada Clinical Trials Group (2005) Erlotinib in previouslytreated non-small-cell lung cancer. N. Engl. J. Med. 353, 123–132

58. Demetri, G. D., von Mehren, M., Blanke, C. D., Van den Abbeele, A. D.,Eisenberg, B., Roberts, P. J., Heinrich, M. C., Tuveson, D. A., Singer, S.,Janicek, M., Fletcher, J. A., Silverman, S. G., Silberman, S. L., Capdeville,R., Kiese, B., Peng, B., Dimitrijevic, S., Druker, B. J., Corless, C.,Fletcher, C. D. M., and Joensuu, H. (2002) Efficacy and safety of imatinibmesylate in advanced gastrointestinal stromal tumors. N. Engl. J. Med.347, 472–480

59. Fabian, M. A., Biggs, W. H., III, Treiber, D. K., Atteridge, C. E., Azimioara,M. D., Benedetti, M. G., Carter, T. A., Ciceri, P., Edeen, P. T., Floyd, M.,Ford, J. M., Galvin, M., Gerlach, J. L., Grotzfeld, R. M., Herrgard, S.,Insko, D. E., Insko, M. A., Lai, A. G., Lelias, J. M., Mehta, S. A., Milanov,Z. V., Velasco, A. M., Wodicka, L. M., Patel, H. K., Zarrinkar, P. P., andLockhart, D. J. (2005) A small molecule-kinase interaction map for clinicalkinase inhibitors. Nat. Biotechnol. 23, 329–336

60. Karaman, M. W., Herrgard, S., Treiber, D. K., Gallant, P., Atteridge, C. E.,Campbell, B. T., Chan, K. W., Ciceri, P., Davis, M. I., Edeen, P. T.,Faraoni, R., Floyd, M., Hunt, J. P., Lockhart, D. J., Milanov, Z. V.,Morrison, M. J., Pallares, G., Patel, H. K., Pritchard, S., Wodicka, L. M.,and Zarrinkar, P. P. (2008) A quantitative analysis of kinase inhibitorselectivity. Nat. Biotechnol. 26, 127–132

61. Hantschel, O., Rix, U., Schmidt, U., Burckstummer, T., Kneidinger, M.,Schutze, G., Colinge, J., Bennett, K. L., Ellmeier, W., Valent, P., andSuperti-Furga, G. (2007) The Btk tyrosine kinase is a major target of theBcr-Abl inhibitor dasatinib. Proc. Natl. Acad. Sci. U. S. A. 104,13283–13288

62. Rix, U., Hantschel, O., Durnberger, G., Remsing Rix, L. L., Planyavsky, M.,Fernbach, N. V., Kaupe, I., Bennett, K. L., Valent, P., Colinge, J., Kocher,T., and Superti-Furga, G. (2007) Chemical proteomic profiles of theBCR-ABL inhibitors imatinib, nilotinib, and dasatinib reveal novel kinaseand nonkinase targets. Blood 110, 4055–4063

63. Bild, A. H., Yao, G., Chang, J. T., Wang, Q., Potti, A., Chasse, D., Joshi,M. B., Harpole, D., Lancaster, J. M., Berchuck, A., Olson, J. A., Jr.,Marks, J. R., Dressman, H. K., West, M., and Nevins, J. R. (2006)Oncogenic pathway signatures in human cancers as a guide to targetedtherapies. Nature 439, 353–357

64. Lynch, T. J., Adjei, A. A., Bunn, P. A., Jr., Eisen, T. G., Engelman, J., Goss,G. D., Haber, D. A., Heymach, J. V., Janne, P. A., Johnson, B. E.,Johnson, D. H., Lilenbaum, R. C., Meyerson, M., Sandler, A. B., Sequist,L. V., Settleman, J., Wong, K. K., and Hart, C. S. (2006) Summarystatement: novel agents in the treatment of lung cancer: advances in



epidermal growth factor receptor-targeted agents. Clin. Cancer Res. 12,4365s–4371s

65. Jensen, O. N. (2007) Interpreting the protein language using proteomics.Nat. Rev. Mol. Cell Biol. 7, 391–403

66. Pawson, T., and Scott, J. D. (1997) Signaling through scaffold, anchoring,and adaptor proteins. Science 278, 2075–2080

67. Bose, R., Molina, H., Patterson, A. S., Bitok, J. K., Periaswamy, B., Bader,J. S., Pandey, A., and Cole, P. A. (2006) Phosphoproteomic analysis ofHer2/neu signaling and inhibition. Proc. Natl. Acad. Sci. U. S. A. 103,9773–9778

68. Goss, V. L., Lee, K. A., Moritz, A., Nardone, J., Spek, E. J., MacNeill, J.,Rush, J., Comb, M. J., and Polakiewicz, R. D. (2006) A common phos-photyrosine signature for the Bcr-Abl kinase. Blood 107, 4888–4897

69. Rush, J., Moritz, A., Lee, K. A., Guo, A., Goss, V. L., Spek, E. J., Zhang, H.,Zha, X. M., Polakiewicz, R. D., and Comb, M. J. (2005) Immunoaffinityprofiling of tyrosine phosphorylation in cancer cells. Nat. Biotechnol. 23,94–101

70. Walters, D. K., Goss, V. L., Stoffregen, E. P., Gu, T. L., Lee, K., Nardone, J.,McGreevey, L., Heinrich, M. C., Deininger, M. W., Polakiewicz, R., andDruker, B. J. (2006) Phosphoproteomic analysis of AML cell lines iden-tifies leukemic oncogenes. Leuk. Res. 30, 1097–1104

71. Walters, D. K., Mercher, T., Gu, T. L., O’Hare, T., Tyner, J. W., Loriaux, M.,Goss, V. L., Lee, K. A., Eide, C. A., Wong, M. J., Stoffregen, E. P.,McGreevey, L., Nardone, J., Moore, S. A., Crispino, J., Boggon, T. J.,Heinrich, M. C., Deininger, M. W., Polakiewicz, R. D., Gilliland, D. G., andDruker, B. J. (2006) Activating alleles of JAK3 in acute megakaryoblasticleukemia. Cancer Cell 10, 65–75

72. Rikova, K., Guo, A., Zeng, Q., Possemato, A., Yu, J., Haack, H., Nardone,J., Lee, K., Reeves, C., Li, Y., Hu, Y., Tan, Z., Stokes, M., Sullivan, L.,Mitchell, J., Wetzel, R., Macneill, J., Ren, J. M., Yuan, J., Bakalarski,C. E., Villen, J., Kornhauser, J. M., Smith, B., Li, D., Zhou, X., Gygi, S. P.,Gu, T. L., Polakiewicz, R. D., Rush, J., and Comb, M. J. (2007) Globalsurvey of phosphotyrosine signaling identifies oncogenic kinases in lungcancer. Cell 131, 1190–1203

73. Dierck, K., Machida, K., Voigt, A., Thimm, J., Horstmann, M., Fiedler, W.,Mayer, B. J., and Nollau, P. (2006) Quantitative multiplexed profiling ofcellular signaling networks using phosphotyrosine-specific DNA-taggedSH2 domains. Nat. Methods 3, 737–744

74. Zhang, Y., Wolf-Yadlin, A., and White, F. M. (2007) Quantitative proteomicanalysis of phosphotyrosine-mediated cellular signaling networks. Meth-ods Mol. Biol. 359, 203–212

75. Blagoev, B., Ong, S. E., Kratchmarova, I., and Mann, M. (2004) Temporalanalysis of phosphotyrosine-dependent signaling networks by quantita-tive proteomics. Nat. Biotechnol. 22, 1139–1145

76. Wolf-Yadlin, A., Hautaniemi, S., Lauffenburger, D. A., and White, F. M.(2007) Multiple reaction monitoring for robust quantitative proteomicanalysis of cellular signaling networks. Proc. Natl. Acad. Sci. U. S. A.

104, 5860–586577. Ding, S. J., Qian, W. J., and Smith, R. D. (2007) Quantitative proteomic

approaches for studying phosphotyrosine signaling. Expert Rev. Pro-teomics 4, 13–23

78. Krutzik, P. O., Crane, J. M., Clutter, M. R., and Nolan, G. P. (2008) High-content single-cell drug screening with phosphospecific flow cytometry.Nat. Chem. Biol. 4, 132–142

79. O’Neill, R. A., Bhamidipati, A., Bi, X., Deb-Basu, D., Cahill, L., Ferrante, J.,Gentalen, E., Glazer, M., Gossett, J., Hacker, K., Kirby, C., Knittle, J.,Loder, R., Mastroieni, C., Maclaren, M., Mills, T., Nguyen, U., Parker, N.,Rice, A., Roach, D., Suich, D., Voehringer, D., Voss, K., Yang, J., Yang,T., and Vander Horn, P. B. (2006) Isoelectric focusing technology quan-tifies protein signaling in 25 cells. Proc. Natl. Acad. Sci. U. S. A. 103,16153–16158

80. Kyle, R. A., Therneau, T. M., Rajkumar, S. V., Larson, D. R., Plevak, M. F.,Offord, J. R., Dispenzieri, A., Katzmann, J. A., and Melton, L. J., III (2006)Prevalence of monoclonal gammopathy of undetermined significance.N. Engl. J. Med. 354, 1362–1369

81. Durie, B. G., and Salmon, S. E. (1975) A clinical staging system for multiplemyeloma. Correlation of measured myeloma cell mass with presentingclinical features, response to treatment, and survival. Cancer 36,842–854

82. Greipp, P. R., San Miguel, J., Durie, B. G., Crowley, J. J., Barlogie, B.,Blade, J., Boccadoro, M., Child, J. A., Avet-Loiseau, H., Kyle, R. A.,Lahuerta, J. J., Ludwig, H., Morgan, G., Powles, R., Shimizu, K., Shustik,C., Sonneveld, P., Tosi, P., Turesson, I., and Westin, J. (2005) Interna-tional staging system for multiple myeloma. J. Clin. Oncol. 23,3412–3420

83. Voorhees, P. M., and Orlowski, R. Z. (2007) Emerging data on the use ofanthracyclines in combination with Bortezomib in multiple myeloma.Clin. Lymphoma Myeloma 7, S156–S162

84. Manochakian, R., Miller, K. C., and Chahan-Khan, A. A. (2007) ClinicalImpact of Bortezomib in frontline regimens for patients with multiplemyeloma. Oncologist 12, 978–990

85. Meister, S., Schubert, U., Neubert, K., Hermann, K., Burger, R., Gramatzki,M., Hahn, S., Schrieber, S., Wilhelm, S., Herrmann, M., Jaeck, H.-M., andVoll, R. E. (2007) Extensive immunoglobulin production sensitizes my-eloma cells for proteasome inhibition. Cancer Res. 67, 1783–1792

86. Simon, G., Sharma, A., Li, X., Hazelton, T., Walsh, F., Williams, C., Chiap-pori, A., Haura, E., Tanvetyanon, T., Antonia, S., Cantor, A., and Bepler,G. (2007) Feasibility and efficacy of molecular analysis-directed individ-ualized therapy in advanced non-small-cell lung cancer. J. Clin. Oncol.25, 2741–2746

87. Zheng, Z., Chen, T., Li, X., Haura, E., Sharma, A., and Bepler, G. (2007) DNAsynthesis and repair genes RRM1 and ERCC1 in lung cancer. N. Engl.J. Med. 356, 800–808



Documents

Proteomic Contributions to Personalized Cancer Care