Artigo Acesso aberto Revisado por pares

Phosphoproteomic Analysis of Signaling Pathways in Head and Neck Squamous Cell Carcinoma Patient Samples

2011; Elsevier BV; Volume: 178; Issue: 2 Linguagem: Inglês

10.1016/j.ajpath.2010.10.044

ISSN

1525-2191

Autores

Mitchell J. Frederick, Amy J. VanMeter, Mayur A. Gadhikar, Ying C. Henderson, Hui Yao, Curtis R. Pickering, Michelle D. Williams, Adel K. El‐Naggar, Vlad C. Sandulache, Emily Tarco, Jeffrey N. Myers, Gary L. Clayman, Lance A. Liotta, Emanuel F. Petricoin, Valerie Calvert, Valentina Fodale, Jing Wang, Randal S. Weber,

Tópico(s)

Monoclonal and Polyclonal Antibodies Research

Resumo

Molecular targeted therapy represents a promising new strategy for treating cancers because many small-molecule inhibitors targeting protein kinases have recently become available. Reverse-phase protein microarrays (RPPAs) are a useful platform for identifying dysregulated signaling pathways in tumors and can provide insight into patient-specific differences. In the present study, RPPAs were used to examine 60 protein end points (predominantly phosphoproteins) in matched tumor and nonmalignant biopsy specimens from 23 patients with head and neck squamous cell carcinoma to characterize the cancer phosphoproteome. RPPA identified 18 of 60 analytes globally elevated in tumors versus healthy tissue and 17 of 60 analytes that were decreased. The most significantly elevated analytes in tumor were checkpoint kinase (Chk) 1 serine 345 (S345), Chk 2 S33/35, eukaryotic translation initiation factor 4E-binding protein 1 (4E-BP1) S65, protein kinase C (PKC) ζ/ι threonine 410/412 (T410/T412), LKB1 S334, inhibitor of kappaB alpha (IκB-α) S32, eukaryotic translation initiation factor 4E (eIF4E) S209, Smad2 S465/67, insulin receptor substrate 1 (IRS-1) S612, mitogen-activated ERK kinase 1/2 (MEK1/2) S217/221, and total PKC ι. To our knowledge, this is the first report of elevated PKC ι in head and neck squamous cell carcinoma that may have significance because PKC ι is an oncogene in several other tumor types, including lung cancer. The feasibility of using RPPA for developing theranostic tests to guide personalized therapy is discussed in the context of these data. Molecular targeted therapy represents a promising new strategy for treating cancers because many small-molecule inhibitors targeting protein kinases have recently become available. Reverse-phase protein microarrays (RPPAs) are a useful platform for identifying dysregulated signaling pathways in tumors and can provide insight into patient-specific differences. In the present study, RPPAs were used to examine 60 protein end points (predominantly phosphoproteins) in matched tumor and nonmalignant biopsy specimens from 23 patients with head and neck squamous cell carcinoma to characterize the cancer phosphoproteome. RPPA identified 18 of 60 analytes globally elevated in tumors versus healthy tissue and 17 of 60 analytes that were decreased. The most significantly elevated analytes in tumor were checkpoint kinase (Chk) 1 serine 345 (S345), Chk 2 S33/35, eukaryotic translation initiation factor 4E-binding protein 1 (4E-BP1) S65, protein kinase C (PKC) ζ/ι threonine 410/412 (T410/T412), LKB1 S334, inhibitor of kappaB alpha (IκB-α) S32, eukaryotic translation initiation factor 4E (eIF4E) S209, Smad2 S465/67, insulin receptor substrate 1 (IRS-1) S612, mitogen-activated ERK kinase 1/2 (MEK1/2) S217/221, and total PKC ι. To our knowledge, this is the first report of elevated PKC ι in head and neck squamous cell carcinoma that may have significance because PKC ι is an oncogene in several other tumor types, including lung cancer. The feasibility of using RPPA for developing theranostic tests to guide personalized therapy is discussed in the context of these data. Head and neck squamous cell carcinoma (HNSCC) is the seventh leading cause of cancer death worldwide and globally afflicts approximately 500,000 new persons each year.1Parkin D.M. Bray F. Ferlay J. Pisani P. Estimating the world cancer burden: globocan 2000.Int J Cancer. 2001; 94: 153-156Crossref PubMed Scopus (3303) Google Scholar In 2009, approximately 48,000 new cases of this devastating and often fatal disease were diagnosed in the United States alone.2Jemal A. Siegel R. Ward E. Hao Y. Xu J. Thun M.J. Cancer statistics, 2009.CA Cancer J Clin. 2009; 59: 225-249Crossref PubMed Scopus (9886) Google Scholar Despite advances in multimodality treatments, the overall 5-year survival rate for patients with HNSCC is only 61% (National Cancer Institute Surveillance, Epidemiology, and End Results database at http://seer.cancer.gov/). The need for new treatment strategies is obvious. The major challenge in developing new therapies is identifying agents that kill cancer cells but do not harm patients. Recognition that tumors are “addicted” to their aberrant pathways,3Sharma S.V. Settleman J. Oncogene addiction: setting the stage for molecularly targeted cancer therapy.Genes Dev. 2007; 21: 3214-3231Crossref PubMed Scopus (343) Google Scholar together with prolific discovery of low-molecular-weight inhibitors that interfere with these pathways, has spawned a virtual explosion of new targeted therapies for treating cancer. Protein kinases have emerged as an important class of therapeutic targets partly because of the ease with which small-molecule inhibitors can be designed. More important, many defining characteristics of cancer, including autocrine growth, insensitivity to antigrowth signals, and evasion of programmed cell death, are heavily regulated by protein phosphorylation.4Dancey J. Sausville E.A. Issues and progress with protein kinase inhibitors for cancer treatment.Nat Rev Drug Discov. 2003; 2: 296-313Crossref PubMed Scopus (489) Google Scholar Because there are 530 different human protein kinases acting in balance with approximately 180 phosphatases, essential questions arise as to which of these kinases (or phosphatases) and their plethora of downstream substrates will be effective cancer targets and in which subset of patients. Several aberrant signaling pathways, along with their activating mechanisms and downstream substrates, have been previously characterized in HNSCC. Among these pathways, the epidermal growth factor receptor (EGFR) has been the most widely studied. Overexpression of EGFR protein occurs in approximately 70% to 80% of HNSCC tumors,5Rubin G.J. Melhem M.F. Barnes E.L. Tweardy D.J. Quantitative immunohistochemical analysis of transforming growth factor-alpha and epidermal growth factor receptor in patients with squamous cell carcinoma of the head and neck.Cancer. 1996; 78: 1284-1292Crossref PubMed Scopus (48) Google Scholar, 6Putti T.C. To K.F. Hsu H.C. Chan A.T. Lai G.M. Tse G. Lee Y.S. Whang-Peng J. Millward M. Lin L. Lin X. Lee C.S. Expression of epidermal growth factor receptor in head and neck cancers correlates with clinical progression: a multicentre immunohistochemical study in the Asia-Pacific region.Histopathology. 2002; 41: 144-151Crossref PubMed Scopus (66) Google Scholar, 7Sheikh Ali M.A. Gunduz M. Nagatsuka H. Gunduz E. Cengiz B. Fukushima K. Beder L.B. Demircan K. Fujii M. Yamanaka N. Shimizu K. Grenman R. Nagai N. Expression and mutation analysis of epidermal growth factor receptor in head and neck squamous cell carcinoma.Cancer Sci. 2008; 99: 1589-1594Crossref PubMed Scopus (94) Google Scholar the gene encoding EGFR is amplified in 17% to 31% of patient tumors,8Temam S. Kawaguchi H. El-Naggar A.K. Jelinek J. Tang H. Liu D.D. Lang W. Issa J.P. Lee J.J. Mao L. Epidermal growth factor receptor copy number alterations correlate with poor clinical outcome in patients with head and neck squamous cancer.J Clin Oncol. 2007; 25: 2164-2170Crossref PubMed Scopus (347) Google Scholar, 9Mrhalova M. Plzak J. Betka J. Kodet R. Epidermal growth factor receptor: its expression and copy numbers of EGFR gene in patients with head and neck squamous cell carcinomas.Neoplasma. 2005; 52: 338-343PubMed Google Scholar, 10Chung C.H. Ely K. McGavran L. Varella-Garcia M. Parker J. Parker N. Jarrett C. Carter J. Murphy B.A. Netterville J. Burkey B.B. Sinard R. Cmelak A. Levy S. Yarbrough W.G. Slebos R.J. Hirsch F.R. Increased epidermal growth factor receptor gene copy number is associated with poor prognosis in head and neck squamous cell carcinomas.J Clin Oncol. 2006; 24: 4170-4176Crossref PubMed Scopus (486) Google Scholar and high levels of the receptor have been correlated with poor prognosis.8Temam S. Kawaguchi H. El-Naggar A.K. Jelinek J. Tang H. Liu D.D. Lang W. Issa J.P. Lee J.J. Mao L. Epidermal growth factor receptor copy number alterations correlate with poor clinical outcome in patients with head and neck squamous cancer.J Clin Oncol. 2007; 25: 2164-2170Crossref PubMed Scopus (347) Google Scholar, 10Chung C.H. Ely K. McGavran L. Varella-Garcia M. Parker J. Parker N. Jarrett C. Carter J. Murphy B.A. Netterville J. Burkey B.B. Sinard R. Cmelak A. Levy S. Yarbrough W.G. Slebos R.J. Hirsch F.R. Increased epidermal growth factor receptor gene copy number is associated with poor prognosis in head and neck squamous cell carcinomas.J Clin Oncol. 2006; 24: 4170-4176Crossref PubMed Scopus (486) Google Scholar, 11Pomerantz R.G. Grandis J.R. The role of epidermal growth factor receptor in head and neck squamous cell carcinoma.Curr Oncol Rep. 2003; 5: 140-146Crossref PubMed Scopus (52) Google Scholar Abnormal signaling through the EGFR has been implicated in the activation of downstream pathways [ie, phosphatidylinositol 3-kinase/Akt and Ras/Raf/mitogen-activated ERK kinase (MEK)/extracellular signal-regulated kinase (ERK)] in HNSCC.12Molinolo A.A. Amornphimoltham P. Squarize C.H. Castilho R.M. Patel V. Gutkind J.S. Dysregulated molecular networks in head and neck carcinogenesis.Oral Oncol. 2009; 45: 324-334Abstract Full Text Full Text PDF PubMed Scopus (289) Google Scholar Paradoxically, few patients with HNSCC respond to drugs targeting the EGFR,13Brockstein B. Lacouture M. Agulnik M. The role of inhibitors of the epidermal growth factor in management of head and neck cancer.J Natl Compr Canc Netw. 2008; 6: 696-706PubMed Google Scholar, 14Mehra R. Cohen R.B. Burtness B.A. The role of cetuximab for the treatment of squamous cell carcinoma of the head and neck.Clin Adv Hematol Oncol. 2008; 6: 742-750PubMed Google Scholar, 15Loeffler-Ragg J. Schwentner I. Sprinzl G.M. Zwierzina H. EGFR inhibition as a therapy for head and neck squamous cell carcinoma.Expert Opin Investig Drugs. 2008; 17: 1517-1531Crossref PubMed Scopus (41) Google Scholar raising questions regarding how addicted most of these tumors are to the EGFR signaling pathway. Indeed, it has been hypothesized that other growth factor receptors, such as insulin-like growth factor 1 receptor,16Slomiany M.G. Black L.A. Kibbey M.M. Tingler M.A. Day T.A. Rosenzweig S.A. Insulin-like growth factor-1 receptor and ligand targeting in head and neck squamous cell carcinoma.Cancer Lett. 2007; 248: 269-279Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar, 17Brady G. Crean S.J. Naik P. Kapas S. Upregulation of IGF-2 and IGF-1 receptor expression in oral cancer cell lines.Int J Oncol. 2007; 31: 875-881PubMed Google Scholar can serve to activate the same downstream pathways in HNSCC in a redundant fashion. Aberrant or constitutive activation of transcription factor–signaling molecules [eg, signal transducer and activator of transcription 3 (Stat3)18Grandis J.R. Drenning S.D. Zeng Q. Watkins S.C. Melhem M.F. Endo S. Johnson D.E. Huang L. He Y. Kim J.D. Constitutive activation of Stat3 signaling abrogates apoptosis in squamous cell carcinogenesis in vivo.Proc Natl Acad Sci U S A. 2000; 97: 4227-4232Crossref PubMed Scopus (564) Google Scholar, 19Nagpal J.K. Mishra R. Das B.R. Activation of Stat-3 as one of the early events in tobacco chewing-mediated oral carcinogenesis.Cancer. 2002; 94: 2393-2400Crossref PubMed Scopus (103) Google Scholar, 20Masuda M. Suzui M. Yasumatu R. Nakashima T. Kuratomi Y. Azuma K. Tomita K. Komiyama S. Weinstein I.B. Constitutive activation of signal transducers and activators of transcription 3 correlates with cyclin D1 overexpression and may provide a novel prognostic marker in head and neck squamous cell carcinoma.Cancer Res. 2002; 62: 3351-3355PubMed Google Scholar and nuclear factor (NF)-κB21Ondrey F.G. Dong G. Sunwoo J. Chen Z. Wolf J.S. Crowl-Bancroft C.V. Mukaida N. Van Waes C. Constitutive activation of transcription factors NF-(kappa)B, AP-1, and NF-IL6 in human head and neck squamous cell carcinoma cell lines that express pro-inflammatory and pro-angiogenic cytokines.Mol Carcinog. 1999; 26: 119-129Crossref PubMed Scopus (211) Google Scholar, 22Tamatani T. Azuma M. Aota K. Yamashita T. Bando T. Sato M. Enhanced IkappaB kinase activity is responsible for the augmented activity of NF-kappaB in human head and neck carcinoma cells.Cancer Lett. 2001; 171: 165-172Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar, 23Nakayama H. Ikebe T. Beppu M. Shirasuna K. High expression levels of nuclear factor kappaB: IkappaB kinase alpha and Akt kinase in squamous cell carcinoma of the oral cavity.Cancer. 2001; 92: 3037-3044Crossref PubMed Scopus (148) Google Scholar] has also been extensively reported in both established cell lines and tumor specimens derived from patients with HNSCC. Both Stat3 and NF-κB are highly regulated by protein kinases. The activation of Stat3 is associated with increased tumor cell proliferation and anti-apoptotic functions.24Leeman R.J. Lui V.W. Grandis J.R. STAT3 as a therapeutic target in head and neck cancer.Expert Opin Biol Ther. 2006; 6: 231-241Crossref PubMed Scopus (204) Google Scholar, 25Lun M. Zhang P.L. Pellitteri P.K. Law A. Kennedy T.L. Brown R.E. Nuclear factor-kappaB pathway as a therapeutic target in head and neck squamous cell carcinoma: pharmaceutical and molecular validation in human cell lines using Velcade and siRNA/NF-kappaB.Ann Clin Lab Sci. 2005; 35: 251-258PubMed Google Scholar, 26Aggarwal S. Takada Y. Singh S. Myers J.N. Aggarwal B.B. Inhibition of growth and survival of human head and neck squamous cell carcinoma cells by curcumin via modulation of nuclear factor-kappaB signaling.Int J Cancer. 2004; 111: 679-692Crossref PubMed Scopus (247) Google Scholar The activation of Akt in primary HNSCC tumors is controversial because some groups27Amornphimoltham P. Sriuranpong V. Patel V. Benavides F. Conti C.J. Sauk J. Sausville E.A. Molinolo A.A. Gutkind J.S. Persistent activation of the Akt pathway in head and neck squamous cell carcinoma: a potential target for UCN-01.Clin Cancer Res. 2004; 10: 4029-4037Crossref PubMed Scopus (148) Google Scholar, 28Mandal M. Younes M. Swan E.A. Jasser S.A. Doan D. Yigitbasi O. McMurphey A. Ludwick J. El-Naggar A.K. Bucana C. Mills G.B. Myers J.N. The Akt inhibitor KP372-1 inhibits proliferation and induces apoptosis and anoikis in squamous cell carcinoma of the head and neck.Oral Oncol. 2006; 42: 430-439Abstract Full Text Full Text PDF PubMed Scopus (57) Google Scholar have reported cancer-specific increases in Akt phosphorylation and others have found little difference between tumor and normal29Pedrero J.M. Carracedo D.G. Pinto C.M. Zapatero A.H. Rodrigo J.P. Nieto C.S. Gonzalez M.V. Frequent genetic and biochemical alterations of the PI 3-K/AKT/PTEN pathway in head and neck squamous cell carcinoma.Int J Cancer. 2005; 114: 242-248Crossref PubMed Scopus (223) Google Scholar or even elevated phosphorylation of Akt in adjacent nonmalignant mucosa from patients.30Tosi L. Rinaldi E. Carinci F. Farina A. Pastore A. Pelucchi S. Cassano L. Evangelisti R. Carinci P. Volinia S. Akt, protein kinase C, and mitogen-activated protein kinase phosphorylation status in head and neck squamous cell carcinoma.Head Neck. 2005; 27: 130-137Crossref PubMed Scopus (18) Google Scholar Nevertheless, high levels of phosphorylated Akt protein are commonly observed in many established HNSCC lines in vitro, where it most likely functions to regulate cell growth and survival.31Moral M. Paramio J.M. Akt pathway as a target for therapeutic intervention in HNSCC.Histol Histopathol. 2008; 23: 1269-1278PubMed Google Scholar Given the critical role of abnormal protein expression and activation in malignant transformation, several groups32Roesch-Ely M. Nees M. Karsai S. Ruess A. Bogumil R. Warnken U. Schnolzer M. Dietz A. Plinkert P.K. Hofele C. Bosch F.X. Proteomic analysis reveals successive aberrations in protein expression from healthy mucosa to invasive head and neck cancer.Oncogene. 2007; 26: 54-64Crossref PubMed Scopus (104) Google Scholar, 33Baker H. Patel V. Molinolo A.A. Shillitoe E.J. Ensley J.F. Yoo G.H. Meneses-Garcia A. Myers J.N. El-Naggar A.K. Gutkind J.S. Hancock W.S. Proteome-wide analysis of head and neck squamous cell carcinomas using laser-capture microdissection and tandem mass spectrometry.Oral Oncol. 2005; 41: 183-199Abstract Full Text Full Text PDF PubMed Scopus (74) Google Scholar, 34Patel V. Hood B.L. Molinolo A.A. Lee N.H. Conrads T.P. Braisted J.C. Krizman D.B. Veenstra T.D. Gutkind J.S. Proteomic analysis of laser-captured paraffin-embedded tissues: a molecular portrait of head and neck cancer progression.Clin Cancer Res. 2008; 14: 1002-1014Crossref PubMed Scopus (164) Google Scholar have attempted global analysis of the HNSCC proteome using high-throughput platforms based on reverse-phase liquid chromatography and mass spectrometry. Although these studies have identified proteins that appear to be differentially expressed in tumors compared with normal tissue, they lacked the capacity to discriminate changes in protein phosphorylation levels. Consequently, we decided to compare and identify differences in phosphoproteins regulating survival or cell death in primary tumors and normal (ie, nonmalignant) specimens from patients with HNSCC using reverse-phase protein microarrays (RPPAs). Phosphorylation and/or expression levels of 60 different protein targets (end points), predominantly phosphoproteins, were simultaneously measured in a cohort of matched laser-capture microdissected (LCM) tumor and nonmalignant samples from 23 patients with HNSCC. The data were analyzed to identify differences in signaling pathways between HNSCC tumors and nonmalignant tissue. In addition, the feasibility of measuring and profiling alterations unique for an individual patient's tumor was examined. Bioinformatics data of this kind could prove pivotal to discovering novel therapeutic targets and developing a personalized medicine approach to treating cancer. Phosphospecific antibodies used in RPPA and Western blot experiments were from Cell Signaling Technology (Danvers, MA). Isoform-specific antibodies to total protein kinase C (PKC) ι were purchased from Cell Signaling Technology (Ab 2998) and Santa Cruz Biotechnology (Santa Cruz, CA) (H-76 and N-20); the PKC ζ-specific antibody was also from Cell Signaling Technology (AB 9372). Purified recombinant PKC ι and ζ made in baculovirus were obtained from Cell Sciences (Canton, MA). Short tandem repeat profiles for all HNSCC and papillary cell lines used in this study were generated from genomic DNA samples submitted to the Johns Hopkins University Core Fragment Analysis Facility (Baltimore, MD). The authenticity and uniqueness of cell lines were confirmed by cross-comparing their short tandem repeat profiles with the database of more than 600 commonly used cell lines deposited at the American Type Tissue Culture Repository (Manassas, VA). Tumor and adjacent nonmalignant mucosa samples excised from patients with HNSCC undergoing surgery as their primary treatment were snap frozen in the operating room immediately after devascularization. Samples obtained from 30 patients were stored in liquid nitrogen before embedding in optimal cutting temperature compound and sectioning onto glass slides with a cryostat. The slides (7 μm) were fixed in 70% ethanol; washed in purified water (Milli-Q); stained with Mayer's hematoxylin, followed by Scott's tap water substitute; and dehydrated by successive washes with increasing concentrations of ethanol, followed by final dehydration in xylene. Proteinase inhibitors (Roche, Indianapolis, IN) were added to the water and 70% ethanol to minimize protein degradation. Approximately 20,000 cells from each sample were captured onto caps (CapSure Macro LCM; Arcturus, Life Technologies, Carlsbad, CA) using a laser capture microscope (PixCell II; Arcturus, Life Technologies), and captured cells were stored at −80°C before extracting protein with a lysis buffer containing a 1:1 dilution of tissue protein extraction reagent (Pierce Biochemicals, Rockford, IL) and ×2 SDS loading buffer (Invitrogen, Carlsbad, CA) supplemented with 2.5% mercaptoethanol. The elapsed time between staining and freezing captured cells did not exceed 35 minutes for any sample, and 25 tumors and 23 matched control specimens had adequate material for LCM. An additional slide from each sample was stained with a modified procedure that included eosin so that a pathologist (M.D.W.) could review the histological features. All human samples were collected and processed in accordance with approved Institutional Review Board tissue banking and use protocols, and all patients gave written informed consent. Printing of lysates onto nitrocellulose arrays was performed as previously described.35Sheehan K.M. Gulmann C. Eichler G.S. Weinstein J.N. Barrett H.L. Kay E.W. Conroy R.M. Liotta L.A. Petricoin III, E.F. Sheehan K.M. Gulmann C. Eichler G.S. Weinstein J.N. Barrett H.L. Kay E.W. Conroy R.M. Liotta L.A. Petricoin E.F. Signal pathway profiling of epithelial and stromal compartments of colonic carcinoma reveals epithelial-mesenchymal transition.Oncogene. 2008; 27: 323-331Crossref PubMed Scopus (54) Google Scholar Samples were printed in duplicate, using a four-point dilution scheme. High and low control lysates (A431, with or without EGF and HeLa cells and with or without pervanadate) were printed on every slide as an internal control. Microarrays were immunostained as previously described36Wulfkuhle J.D. Speer R. Pierobon M. Laird J. Espina V. Deng J. Mammano E. Yang S.X. Swain S.M. Nitti D. Esserman L.J. Belluco C. Liotta L.A. Petricoin III, E.F. Multiplexed cell signaling analysis of human breast cancer applications for personalized therapy.J Proteome Res. 2008; 7: 1508-1517Crossref PubMed Scopus (120) Google Scholar on an automated slide stainer (Dako, Carpinteria, CA) using a biotin-linked catalyzed signal amplification system (Dako). Stained slides were scanned using an illustrating system (Adobe Photoshop 6.0; Adobe, San Jose, CA) on a scanner (UMAX PowerLook III; UMAX, Dallas, TX) as 16-bit tagged information file format images at 600 dpi. Antibodies used in these studies were validated for specificity by immunoblotting before use on arrays. All end points tested in these studies are listed in Supplemental Table S1 at http://ajp.amjpathol.org. Total protein values were determined after staining (SYPRO Ruby; Molecular Probes, Eugene, OR) of two representative slides, printed near the middle and end of the arraying process. The TIFF images were analyzed using computer software (MicroVigene; VigeneTech, Boston, MA) to determine the background-corrected intensities for each spot. Differences in sample loading were corrected for by normalizing to total protein values determined from stained slides (SYPRO Ruby). Protein levels for each analyte were calculated using an in-house–developed software package (SuperCurve). Essentially, information provided by the whole dilution series from all samples for a given slide (ie, one analyte at a time) is jointly used to estimate the relative protein concentration for an individual sample. The SuperCurve method takes advantage of the fact that each microarray is probed with a single antibody and, thus, all samples on an array should have similar hybridization behavior and measurements from all spots; and can be used to provide information about baseline and saturation levels and rate of signal increase for each dilution point to construct a SuperCurve. The software package uses a three-parameter logistic model to construct a SuperCurve for each analyte, from which an EC50 in a log2 scale can be calculated individually for each sample on an array. The log2 EC50 is an estimate of the true protein concentration for an individual sample's analyte at some prespecified step within the dilution series (in this case, the median step). Paired t-tests were used to test against the null hypothesis of no differential expression between tumor and normal samples for any given protein. The Benjamini-Hochberg method was applied to control the false-discovery rate (FDR) of 0.05. Software (JMP 8.0.1 analysis software; SAS, Cary, NC.) was used to perform the unsupervised cluster analysis and to generate a heat map representing log2 EC50 values from all sample end points. Patient-specific alterations in tumor samples were defined for each analyte using an algorithm that accounted for both the magnitude of an end point and the specific ratio when compared with the patient-matched nonmalignant control. For each end point, cutoff values for magnitude were determined by taking the mean log2 EC50 value for all nonmalignant samples ±1.65 SDs (ie, 90% of measurements expected to fall in between). Cutoff values for ratios were determined by taking the mean of all possible ratios between every nonmalignant sample for each end point ±1.28 SDs (ie, 80% of ratios expected to fall in between) after a log2 transformation. Analytes were considered to be truly increased or decreased in a patient's tumor only if both the log2 EC50 and the tumor EC50/matched control EC50 ratio were outside the predetermined cutoff values. However, for analyzing the feasibility of future diagnostic tests in the absence of patient-matched controls, a test result was considered positive based solely on the magnitude of a tumor's end point. False-positive results corresponded to patients whose tumors were positively elevated by magnitude alone but failed to exceed the cutoff ratio for activation. The positive predictive value for markers was then defined as follows: true positives/(true positives plus false positives). The sensitivity for each analyte was defined as follows: true negatives/(true negatives plus false positives). Potential correlations between sample clusters and various clinicopathological end points for patient samples were examined with a Fisher's exact test. End points significantly associated with tumor and nonmalignant subsets were identified by class comparison using computer software (BRB Array tools software version 3.8.0) developed by Richard Simon, Ph.D., and the BRB-ArrayTools Development Team. Pearson correlation coefficients derived from every possible pair of end points were obtained using software (JMP 8.0.1). Potential signaling modules were identified after performing an unsupervised cluster analysis with the correlation coefficients and generating a heat map using ArrayTools. Correlation values greater than 0.41 (ie, P < 0.005 for ν = 46) were considered statistically significant. Pathway analysis was performed by searching scientific publications using MEDLINE, the Internet (ie, Google), and software (Ingenuity Pathways Analysis v8.5; Ingenuity Systems, Redwood City, CA). Protein from approximately 10,000 cells collected by LCM was solubilized in sample lysis buffer containing 87-mmol/L Tris (pH 6.8), 75-mmol/L sodium chloride, 1-mmol/L EDTA, 0.5% NP40, 2% SDS, 0.25% deoxycholate, 10% glycerol, 2.5% mercaptoethanol, 10-mmol/L sodium fluoride, 1-mmol/L orthovanadate, 2-mmol/L pyrophosphate, and proteinase inhibitors. Sample protein concentrations were estimated by dot blotting serial dilutions onto nitrocellulose, staining protein (SYPRO Ruby) for blots, and comparing the fluorescent intensity of spots under UV transillumination with a dilution series made of bovine serum albumin standards. Fluorescent spot intensities were quantitated with a gel documentation system (Bio-Rad, Hercules, CA) running software (Quantity One). Equal amounts of protein (1 to 3 μg/lane) were heat denatured, resolved by SDS–polyacrylamide gel electrophoresis, and electrotransferred overnight to nitrocellulose membranes. Membranes were blocked for 30 minutes at room temperature with casein (Vector Laboratories, Burlingame, CA); incubated with primary antibodies (diluted 1:1000 in casein) overnight at 4°C; washed with casein the next day; incubated for 45 minutes at room temperature with horseradish peroxidase conjugated goat anti-rabbit (Cell Signaling Technology) (diluted 1:2000 in casein); washed in PBS/0.05% TW-20; incubated for 15 minutes at room temperature with biotinylated tyramide (Perkin Elmer, Waltham, MA); washed with PBS/TW-20/20% dimethyl sulfoxide and then casein; incubated at room temperature with alkaline phosphatase reagent (Vectastain ABC; Vector Laboratories); diluted 1:250 in casein for 10 minutes; and washed, incubated for 5 minutes with 0.1-mol/L Tris (9.5), and incubated for 5 minutes with substrate (DuoLux AP; Vector Laboratories) before rinsing in 0.1-mol/L Tris and exposing to X-ray film. Sections (5 μm) cut from formalin-fixed archival specimens were deparaffinized and hydrated by successive incubations in xylene, 100% ethanol, and 95% ethanol, followed by rinsing in water. After antigen retrieval, immunohistochemistry (IHC) was performed with an autostainer (Lab Vison autostainer 360). Essentially, slides were blocked, incubated with primary antibodies (60 minutes at room temperature, except for phosphorylated PKC ζ/ι T410/T412, which was at 4°C overnight), and signal detected with a detection system (UltraVision LP) using a horseradish peroxidase polymer and diaminobenzidine substrate, followed by counterstaining with hematoxylin. The following antibodies, dilutions, and antigen retrievals from Cell Signaling Technology were used for IHC: anti-checkpoint kinase (Chk) 2 T68 2661 (1:100/EDTA, pH 8), anti-EGFR Y1173 4407 (1:100/Tris-EDTA, pH 9), anti-MEK1/2 S217/221 2338 (1:50/citrate, pH 6), anti-ERK1/2 T202/Y204 4376 (1:150/citrate, pH 6), anti-ErbB3 Y1289 4791 (1:100/EDTA, pH 8), and anti-PKC ζ/ι T410/T412 9378 (1:50/citrate, pH 6). The anti-Chk1 S345 Ab47318 (1:100/citrate, pH 6) was obtained from Abcam (Cambridge, MA); anti-total PKC ι 610175 (1:100/citrate, pH 6), Becton Dickinson laboratories; and anti-total EGFR sc-03 (1:100/citrate, pH 6), Santa Cruz Biotechnology (San Jose, CA). Staining was scored a

Referência(s)