Earlier work identified specific tumor-promoting abnormalities that are shared between lung cancers and adjacent normal bronchial epithelia. We sought to characterize the yet unknown global molecular and adjacent airway field cancerization (FC) in early-stage non–small cell lung cancer (NSCLC).
Whole-transcriptome expression profiling of resected early-stage (I–IIIA) NSCLC specimens (n = 20) with matched tumors, multiple cytologically controlled normal airways with varying distances from tumors, and uninvolved normal lung tissues (n = 194 samples) was performed using the Affymetrix Human Gene 1.0 ST platform. Mixed-effects models were used to identify differentially expressed genes among groups. Ordinal regression analysis was performed to characterize site-dependent airway expression profiles. All statistical tests were two-sided, except where noted.
We identified differentially expressed gene features (n = 1661) between NSCLCs and airways compared with normal lung tissues, a subset of which (n = 299), after gene set enrichment analysis, statistically significantly (P < .001) distinguished large airways in lung cancer patients from airways in cancer-free smokers. In addition, we identified genes (n = 422) statistically significantly and progressively differentially expressed in airways by distance from tumors that were found to be congruently modulated between NSCLCs and normal lung tissues. Furthermore, LAPTM4B, with statistically significantly increased expression (P < .05) in airways with shorter distance from tumors, was upregulated in human immortalized cells compared with normal bronchial epithelial cells (P < .001) and promoted anchorage-dependent and -independent lung cancer cell growth.
The adjacent airway FC comprises both site-independent profiles as well as gradient and localized airway expression patterns. Profiling of the airway FC may provide new insights into NSCLC oncogenesis and molecular tools for detection of the disease.
DNA methylation is associated with aberrant gene expression in cancer, and has been shown to correlate with therapeutic response and disease prognosis in some types of cancer. We sought to investigate the biological significance of DNA methylation in lung cancer.
We integrated the gene expression profiles and data of gene promoter methylation for a large panel of non-small cell lung cancer cell lines, and identified 578 candidate genes with expression levels that were inversely correlated to the degree of DNA methylation. We found these candidate genes to be differentially methylated in normal lung tissue versus non-small cell lung cancer tumors, and segregated by histologic and tumor subtypes. We used gene set enrichment analysis of the genes ranked by the degree of correlation between gene expression and DNA methylation to identify gene sets involved in cellular migration and metastasis. Our unsupervised hierarchical clustering of the candidate genes segregated cell lines according to the epithelial-to-mesenchymal transition phenotype. Genes related to the epithelial-to-mesenchymal transition, such as AXL, ESRP1, HoxB4, and SPINT1/2, were among the nearly 20% of the candidate genes that were differentially methylated between epithelial and mesenchymal cells. Greater numbers of genes were methylated in the mesenchymal cells and their expressions were upregulated by 5-azacytidine treatment. Methylation of the candidate genes was associated with erlotinib resistance in wild-type EGFR cell lines. The expression profiles of the candidate genes were associated with 8-week disease control in patients with wild-type EGFR who had unresectable non-small cell lung cancer treated with erlotinib, but not in patients treated with sorafenib.
Our results demonstrate that the underlying biology of genes regulated by DNA methylation may have predictive value in lung cancer that can be exploited therapeutically.
Electronic supplementary material
The online version of this article (doi:10.1186/1471-2164-15-1079) contains supplementary material, which is available to authorized users.
DNA methylation; Epithelial-mesenchymal transition; Erlotinib; Lung cancer
Motivation: With the expansion of high-throughput technologies, understanding different kinds of genome-level data is a common task. MicroRNA (miRNA) is increasingly profiled using high-throughput technologies (microarrays or next-generation sequencing). The downstream analysis of miRNA targets can be difficult. Although there are many databases and algorithms to predict miRNA targets, there are few tools to integrate miRNA–gene interaction data into high-throughput genomic analyses.
Results: We present targetHub, a CouchDB database of miRNA–gene interactions. TargetHub provides a programmer-friendly interface to access miRNA targets. The Web site provides RESTful access to miRNA–gene interactions with an assortment of gene and miRNA identifiers. It can be a useful tool to integrate miRNA target interaction data directly into high-throughput bioinformatics analyses.
Availability: TargetHub is available on the web at http://app1.bioinformatics.mdanderson.org/tarhub/_design/basic/index.html.
Investigate the mechanisms of regulation and role associated with EZH2 expression in lung cancer cells.
We investigated the mechanisms of EZH2 expression associated with the vascular endothelial growth factor (VEGF)/VEGF receptor 2 (VEGFR-2) pathway. Furthermore, we sought to determine the role of EZH2 in response of lung adenocarcinoma to platinum-based chemotherapy, as well as the effect of EZH2 depletion on VEGFR-2–targeted therapy in lung adenocarcinoma cell lines. Additionally, we characterized EZH2 expression in lung adenocarcinoma specimens and correlated it with patients’ clinical characteristics.
In this study, we demonstrate that VEGF/VEGFR-2 activation induces expression of EZH2 through the upregulation of E2F3 and HIF-1α, and downregulated expression of miR-101. EZH2 depletion by treatment with 3-deazaneplanocin A and knockdown by siRNA decreased the expression of EZH2 and H3K27me3, increased PARP-C level, reduced cell proliferation and migration, and increased sensitivity of the cells to treatment with cisplatin and carboplatin. Additionally, high EZH2 expression was associated with poor overall survival in patients who received platinum-based adjuvant therapy, but not in patients who did not receive this therapy. Furthermore, we demonstrated for the first time that the inhibition of EZH2 greatly increased the sensitivity of lung adenocarcinoma cells to the anti-VEGFR-2 drug AZD2171.
Our results suggest that VEGF/VEGFR-2 pathway plays a role in regulation of EZH2 expression via E2F3, HIF-1α and miR-101. EZH2 depletion decreases the malignant potential of lung adenocarcinoma and sensitivity of the cells to both platinum-based and VEGFR-2–targeted therapy.
EZH2; NSCLC; VEGF/VEGFR-2 pathway; DZNep
Lysophosphatidic acid (LPA) acts through high affinity G protein-coupled receptors to mediate a plethora of physiological and pathological activities associated with tumorigenesis. LPA receptors and autotaxin (ATX/LysoPLD), the primary enzyme producing LPA, are aberrantly expressed in multiple cancer lineages. However, the role of ATX and LPA receptors in the initiation and progression of breast cancer has not been evaluated. We demonstrate that expression of ATX or each Edg-family LPA receptor in mammary epithelium of transgenic mice is sufficient to induce a high frequency of late-onset, estrogen receptor (ER) positive, invasive and metastatic mammary cancer. Thus ATX and LPA receptors can contribute to the initiation and progression of breast cancer.
LPA; ATX; Transgenic mouse model; Breast cancer; Metastasis
Intrinsic resistance to agents targeting phosphatidylinositol-3-kinase (PI3K)/AKT pathway is one of the major challenges in cancer treatment with such agents. The objective of this study is to identify the genes or pathways that can be targeted to overcome the resistance of non-small cell lung cancer to the AKT inhibitor, MK2206, which is currently being evaluated in phase I and II clinical trials. Using a genome-wide small interfering RNA (siRNA) library screening and biological characterization we identified that inhibition of Thioredoxin Reductase-1 (TXNRD1), one of the key anti-oxidant enzymes, with siRNAs or its inhibitor, Auranofin, sensitized non-small cell lung cancer cells to MK2206 treatment in vitro and in vivo. We found that simultaneous inhibition of TXNRD1 and AKT pathways induced robust reactive oxygen species (ROS) production, which was involved in c-Jun N-terminal Kinase (JNK, MAPK8) activation and cell apoptosis. Furthermore we found that the synthetic lethality interaction between the TXNRD1 and AKT pathways occurred through the KEAP1/NRF2 cellular antioxidant pathway. Lastly, we found that synthetic lethality induced by TXNRD1 and AKT inhibitors relied on wild type KEAP1 function. Our study indicates that targeting the interaction between AKT and TXNRD1 antioxidant pathways with MK2206 and Auranofin, a FDA approved drug, is a rational strategy to treat lung cancer and that KEAP1 mutation status may offer a predicative biomarker for such combination approaches.
Synthetic lethality; AKT; TXNRD1; MK2206; KEAP1
The treatment of uveal melanoma with proton radiotherapy has provided excellent clinical outcomes. However, contemporary treatment planning systems use simplistic dose algorithms that limit the accuracy of relative dose distributions. Further, absolute predictions of absorbed dose per monitor unit are not yet available in these systems. The purpose of this study was to determine if Monte Carlo methods could predict dose per monitor unit (D/MU) value at the center of a proton spread-out Bragg peak (SOBP) to within 1% on measured values for a variety of treatment fields relevant to ocular proton therapy. The MCNPX Monte Carlo transport code, in combination with realistic models for the ocular beam delivery apparatus and a water phantom, was used to calculate dose distributions and D/MU values, which were verified by the measurements. Measured proton beam data included central-axis depth dose profiles, relative cross-field profiles and absolute D/MU measurements under several combinations of beam penetration ranges and range-modulation widths. The Monte Carlo method predicted D/MU values that agreed with measurement to within 1% and dose profiles that agreed with measurement to within 3% of peak dose or within 0.5 mm distance-to-agreement. Lastly, a demonstration of the clinical utility of this technique included calculations of dose distributions and D/MU values in a realistic model of the human eye. It is possible to predict D/MU values accurately for clinical relevant range-modulated proton beams for ocular therapy using the Monte Carlo method. It is thus feasible to use the Monte Carlo method as a routine absolute dose algorithm for ocular proton therapy.
Motivation: Identification of bimodally expressed genes is an important task, as genes with bimodal expression play important roles in cell differentiation, signalling and disease progression. Several useful algorithms have been developed to identify bimodal genes from microarray data. Currently, no method can deal with data from next-generation sequencing, which is emerging as a replacement technology for microarrays.
Results: We present SIBER (systematic identification of bimodally expressed genes using RNAseq data) for effectively identifying bimodally expressed genes from next-generation RNAseq data. We evaluate several candidate methods for modelling RNAseq count data and compare their performance in identifying bimodal genes through both simulation and real data analysis. We show that the lognormal mixture model performs best in terms of power and robustness under various scenarios. We also compare our method with alternative approaches, including profile analysis using clustering and kurtosis (PACK) and cancer outlier profile analysis (COPA). Our method is robust, powerful, invariant to shifting and scaling, has no blind spots and has a sample-size-free interpretation.
Availability: The R package SIBER is available at the website http://bioinformatics.mdanderson.org/main/OOMPA:Overview.
Supplementary information: Supplementary data are available at Bioinformatics online.
Proteomic profiling has the potential to impact the diagnosis, prognosis, and treatment of various diseases. A number of different proteomic technologies are available that allow us to look at many proteins at once, and all of them yield complex data that raise significant quantitative challenges. Inadequate attention to these quantitative issues can prevent these studies from achieving their desired goals, and can even lead to invalid results. In this chapter, we describe various ways the involvement of statisticians or other quantitative scientists in the study team can contribute to the success of proteomic research, and we outline some of the key statistical principles that should guide the experimental design and analysis of such studies.
blocking; data preprocessing; experimental design; false discovery rate; image processing; mass spectrometry; peak detection; randomization; spot detection; 2D gel electrophoresis; validation
EMT has been associated with metastatic spread and EGFR inhibitor resistance. We developed and validated a robust 76-gene EMT signature using gene expression profiles from four platforms using NSCLC cell lines and patients treated in the BATTLE study.
We conducted an integrated gene expression, proteomic, and drug response analysis using cell lines and tumors from NSCLC patients. A 76-gene EMT signature was developed and validated using gene expression profiles from four microarray platforms of NSCLC cell lines and patients treated in the BATTLE (Biomarker-integrated Approaches of Targeted Therapy for Lung Cancer Elimination) study, and potential therapeutic targets associated with EMT were identified.
Compared with epithelial cells, mesenchymal cells demonstrated significantly greater resistance to EGFR and PI3K/Akt pathway inhibitors, independent of EGFR mutation status, but more sensitivity to certain chemotherapies. Mesenchymal cells also expressed increased levels of the receptor tyrosine kinase Axl and showed a trend towards greater sensitivity to the Axl inhibitor SGI-7079, while the combination of SGI-7079 with erlotinib reversed erlotinib resistance in mesenchymal lines expressing Axl and in a xenograft model of mesenchymal NSCLC. In NSCLC patients, the EMT signature predicted 8-week disease control in patients receiving erlotinib, but not other therapies.
We have developed a robust EMT signature that predicts resistance to EGFR and PI3K/Akt inhibitors, highlights different patterns of drug responsiveness for epithelial and mesenchymal cells, and identifies Axl as a potential therapeutic target for overcoming EGFR inhibitor resistance associated with the mesenchymal phenotype
lung cancer; EMT; EGFR inhibition; PI3K inhibition; Axl
Gene expression alterations in response to cigarette smoke have been characterized in normal-appearing bronchial epithelium of healthy smokers and it has been suggested that adjacent histologically normal tissue display tumor-associated molecular abnormalities. We sought to delineate the spatial and temporal molecular lung field of injury in smoker early stage non-small cell lung cancer (NSCLC) patients (n=19) who were accrued into a surveillance clinical trial for annual follow-up and bronchoscopies within one year after definitive surgery. Bronchial brushings and biopsies were obtained from six different sites in the lung at the time of inclusion in the study and at 12, 24 and 36 months after the first time point. Affymetrix Human Gene 1.0 ST arrays were used for whole-transcript expression profiling of airways (n=391). Microarray analysis identified gene features (n=1165) that were non-uniform by site and differentially expressed between airways adjacent to tumors relative to more distant samples as well as those (n=1395) that were significantly altered with time up to three years. In addition, gene-interaction networks mediated by PI3K and ERK1/2 were modulated in adjacent compared to contralateral airways and the latter network with time. Furthermore, phosphorylated AKT and ERK1/2 immunohistochemical expression were significantly increased with time (nuclear pAKT, p=0.03; cytoplasmic pAKT, p<0.0001; pERK1/2, p=0.02) and elevated in adjacent compared to more distant airways (nuclear pAKT, p=0.04; pERK1/2, p=0.03). This study highlights spatial and temporal cancer-associated expression alterations in the molecular field of injury of early stage NSCLC patients after definitive surgery that warrant further validation in independent studies.
Early stage NSCLC; gene expression profiling; lung airway epithelium; chemoprevention
DNA methylation and BLC-2 are potential therapeutic targets in acute myeloid leukemia (AML). We investigated pharmacologic interaction between the DNA methyltransferase inhibitor 5-azacytidine (5-AZA) and the BCL-2 inhibitor ABT-737. Increased BCL-2 expression determined by reverse phase protein analysis was associated with poor survival in AML patients with unfavorable cytogenetics (n=195). We found that 5-AZA, which itself has modest apoptotic activity, acts synergistically with ABT-737 to induce apoptosis. The 5-AZA/ABT-737 combination enhanced mitochondrial outer membrane permeabilization, as evidenced by effective conformational activation of BAX and Δψm loss. Although absence of p53 limited apoptotic activities of 5-AZA and ABT-737 as single agents, the combination synergistically induced apoptosis independent of p53 expression. 5-AZA down-regulated MCL-1, known to mediate resistance to ABT-737, in a p53-independent manner. The 5-AZA/ABT-737 combination synergistically induced apoptosis in AML cells from 7 of 8 patients. 5-AZA significantly reduced MCL-1 levels in 2 of 3 samples examined. Our data provide a molecular rationale for this combination strategy in AML therapy.
AML; 5-azacytidine; ABT-737; BCL-2; MCL-1; p53
Motivation: Identifying genes altered in cancer plays a crucial role in both understanding the mechanism of carcinogenesis and developing novel therapeutics. It is known that there are various mechanisms of regulation that can lead to gene dysfunction, including copy number change, methylation, abnormal expression, mutation and so on. Nowadays, all these types of alterations can be simultaneously interrogated by different types of assays. Although many methods have been proposed to identify altered genes from a single assay, there is no method that can deal with multiple assays accounting for different alteration types systematically.
Results: In this article, we propose a novel method, integration using item response theory (integIRTy), to identify altered genes by using item response theory that allows integrated analysis of multiple high-throughput assays. When applied to a single assay, the proposed method is more robust and reliable than conventional methods such as Student’s t-test or the Wilcoxon rank-sum test. When used to integrate multiple assays, integIRTy can identify novel-altered genes that cannot be found by looking at individual assay separately. We applied integIRTy to three public cancer datasets (ovarian carcinoma, breast cancer, glioblastoma) for cross-assay type integration which all show encouraging results.
Availability and implementation: The R package integIRTy is available at the web site http://bioinformatics.mdanderson.org/main/OOMPA:Overview.
Supplementary data are available at Bioinformatics online.
Acute myeloid leukemia (AML) is believed to arise from leukemic stem-like cells (LSC) making understanding the biological differences between LSC and normal stem cells (HSC) or common myeloid progenitors (CMP) crucial to understanding AML biology. To determine if protein expression patterns were different in LSC compared to other AML and CD34+ populations, we measured the expression of 121 proteins by Reverse Phase Protein Arrays (RPPA) in 5 purified fractions from AML marrow and blood samples: Bulk (CD3/CD19 depleted), CD34-, CD34+(CMP), CD34+CD38+ and CD34+CD38-(LSC). LSC protein expression differed markedly from Bulk (n=31 cases, 93/121 proteins) and CD34+ cells (n= 30 cases, 88/121 proteins) with 54 proteins being significantly different (31 higher, 23 lower) in LSC than in either Bulk or CD34+ cells. Sixty-seven proteins differed significantly between CD34+ and Bulk blasts (n=69 cases). Protein expression patterns in LSC and CD34+ differed markedly from normal CD34+ cells. LSC were distinct from CD34+ and Bulk cells by principal component and by protein signaling network analysis which confirmed individual protein analysis. Potential targetable submodules in LSC included the proteins PU.1(SP1), P27, Mcl1, HIF1α, cMET, P53, Yap, and phospho-Stats 1, 5 and 6. Protein expression and activation in LSC differs markedly from other blast populations suggesting that studies of AML biology should be performed in LSC.
Small cell lung cancer (SCLC) is an aggressive malignancy distinct from non-small cell lung cancer (NSCLC) in its metastatic potential and treatment response. Using an integrative proteomic and transcriptomic analysis, we investigated molecular differences contributing to the distinct clinical behavior of SCLC and NSCLC. SCLC demonstrated lower levels of several receptor tyrosine kinases and decreased activation of PI3K and Ras/MEK pathways, but significantly increased levels of E2F1-regulated factors including EZH2, thymidylate synthase, apoptosis mediators, and DNA repair proteins. Additionally, poly (ADP-ribose) polymerase 1 (PARP1), a DNA repair protein and E2F1 co-activator, was highly expressed at the mRNA and protein levels in SCLC. SCLC growth was inhibited by PARP1 and EZH2 knockdown. Furthermore, SCLC was significantly more sensitive to PARP inhibitors than NSCLC, and PARP inhibition downregulated key components of the DNA repair machinery and enhanced the efficacy of chemotherapy.
On the basis of reversal of taxane resistance with AKT inhibition, we initiated a phase I trial of the AKT inhibitor perifosine with docetaxel in taxane and platinum-resistant or refractory epithelial ovarian cancer.
Patients with pathologically confirmed high-grade epithelial ovarian cancer (taxane resistant, n = 10; taxane refractory, n = 11) were enrolled. Peripheral blood samples and tumor biopsies were obtained and 18F-FDG-PET and DCE-MRI scans were performed for pharmacodynamic and imaging studies.
Patients received a total of 42 treatment cycles. No dose-limiting toxicity was observed. The median progression-free survival and overall survival were 1.9 months and 4.5 months, respectively. One patient with a PTEN mutation achieved a partial remission (PR) for 7.5 months, and another patient with PIK3CA mutation had stable disease (SD) for 4 months. Two other patients without apparent PI3K pathway aberrations achieved SD. Two patients with RAS mutations demonstrated rapid progression. Decreased phosphorylated S6 correlated with 18F-FDG-PET responses.
Patients tolerated perifosine 150 mg PO daily plus docetaxel at 75 mg/m2 every 4 weeks. Further clinical evaluation of effects of perifosine with docetaxel on biological markers and efficacy in patients with ovarian cancer with defined PI3K pathway mutational status is warranted.
Peroxisome proliferator-activated receptor-γ (PPARγ) is a member of the nuclear receptor family of transcription factors with important regulatory roles in cellular growth, differentiation and apoptosis. Using proteomic analysis, we demonstrated expression of PPARγ protein in a series of 260 newly diagnosed primary AML samples. Forced expression of PPARγ enhanced the sensitivity of myeloid leukemic cells to PPARγ agonists CDDO- or 15d15DPGJ2-induced apoptosis, through preferential cleavage of caspase-8. No effects on cell cycle distribution or differentiation were noted, despite prominent induction of p21 in PPARγ-transfected cells. In turn, antagonizing PPARγ function by siRNA or pharmacological PPARγ inhibitor significantly diminished apoptosis induction by CDDO. Overexpression of co-activator protein DRIP205 resulted in enhanced differentiation induction by CDDO in AML cells through PPARγ activation. Studies with DRIP205 deletion constructs demonstrated that the NR boxes of DRIP205 are not required for this co-activation. In a Phase I clinical trial of CDDO (RTA-401) in leukemia, CDDO induced an increase in PPARγ mRNA expression in 6 of 9 patient samples; of those, induction of differentiation was documented in 4, and of p21 in 3 patients, all expressing DRIP205 protein. In summary, these findings suggest that cellular levels of PPARγ regulate induction of apoptosis via caspase-8 activation, while the co-activator DRIP205 is a determinant of induction of differentiation, in response to PPARγ agonists in leukemic cells.
PPARgamma; DRIP205; AML; CDDO; differentiation; apoptosis
With the proliferation of high-throughput technologies, genome-level data analysis has become common in molecular biology. Bioinformaticians are developing extensive resources to annotate and mine biological features from high-throughput data. The underlying database management systems for most bioinformatics software are based on a relational model. Modern non-relational databases offer an alternative that has flexibility, scalability, and a non-rigid design schema. Moreover, with an accelerated development pace, non-relational databases like CouchDB can be ideal tools to construct bioinformatics utilities. We describe CouchDB by presenting three new bioinformatics resources: (a) geneSmash, which collates data from bioinformatics resources and provides automated gene-centric annotations, (b) drugBase, a database of drug-target interactions with a web interface powered by geneSmash, and (c) HapMap-CN, which provides a web interface to query copy number variations from three SNP-chip HapMap datasets. In addition to the web sites, all three systems can be accessed programmatically via web services.
NoSQL database; copy number variation; drug-target interaction; data integration
Mutations in the v-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog (KRAS) play a critical role in cancer cell growth and resistance to therapy. Most mutations occur at codons 12 and 13. In colorectal cancer, the presence of any mutant KRas amino acid substitution is a negative predictor of patient response to targeted therapy. However, in non–small cell lung cancer (NSCLC), the evidence that KRAS mutation is a predictive factor is conflicting.
We used data from a molecularly targeted clinical trial for 215 patients with tissues available out of 268 evaluable patients with refractory NSCLC to examine associations between specific mutant KRas proteins and progression-free survival and tumor gene expression. Transcriptome microarray studies of patient tumor samples and reverse-phase protein array studies of a panel of 67 NSCLC cell lines with known substitutions in KRas and in immortalized human bronchial epithelial cells stably expressing different mutant KRas proteins were used to investigate signaling pathway activation. Molecular modeling was used to study the conformations of wild-type and mutant KRas proteins. Kaplan–Meier curves and Cox regression were used to analyze survival data. All statistical tests were two-sided.
Patients whose tumors had either mutant KRas-Gly12Cys or mutant KRas-Gly12Val had worse progression-free survival compared with patients whose tumors had other mutant KRas proteins or wild-type KRas (P = .046, median survival = 1.84 months) compared with all other mutant KRas (median survival = 3.35 months) or wild-type KRas (median survival = 1.95 months). NSCLC cell lines with mutant KRas-Gly12Asp had activated phosphatidylinositol 3-kinase (PI-3-K) and mitogen-activated protein/extracellular signal-regulated kinase kinase (MEK) signaling, whereas those with mutant KRas-Gly12Cys or mutant KRas-Gly12Val had activated Ral signaling and decreased growth factor–dependent Akt activation. Molecular modeling studies showed that different conformations imposed by mutant KRas may lead to altered association with downstream signaling transducers.
Not all mutant KRas proteins affect patient survival or downstream signaling in a similar way. The heterogeneous behavior of mutant KRas proteins implies that therapeutic interventions may need to take into account the specific mutant KRas expressed by the tumor.
The requirement of frozen tissues for microarray experiments limits the clinical usage of genome-wide expression profiling using microarray technology. The goal of this study is to test the feasibility of developing lung cancer prognosis gene signatures using genome-wide expression profiling of formalin-fixed paraffin-embedded (FFPE) samples, which are widely available and provide a valuable rich source for studying the association of molecular changes in cancer and associated clinical outcomes.
We randomly selected 100 Non-Small-Cell lung cancer (NSCLC) FFPE samples with annotated clinical information from the UT-Lung SPORE Tissue Bank. We micro dissected tumor area from FFPE specimens, and used Affymetrix U133 plus 2.0 arrays to attain gene expression data. After strict quality control and analysis procedures, a supervised principal component analysis was used to develop a robust prognosis signature for NSCLC. Three independent published microarray data sets were used to validate the prognosis model.
This study demonstrated that the robust gene signature derived from genome-wide expression profiling of FFPE samples is strongly associated with lung cancer clinical outcomes, can be used to refine the prognosis for stage I lung cancer patients and the prognostic signature is independent of clinical variables. This signature was validated in several independent studies and was refined to a 59-gene lung cancer prognosis signature.
We conclude that genome-wide profiling of FFPE lung cancer samples can identify a set of genes whose expression level provides prognostic information across different platforms and studies, which will allow its application in clinical settings.
Lung Cancer Prognosis; Gene Expression Signature; Formalin Fixed Paraffin Embedded Samples
High-throughtput technologies enable the testing of tens of thousands of measurements simultaneously. Identification of genes that are differentially expressed or associated with clinical outcomes invokes the multiple testing problem. False Discovery Rate (FDR) control is a statistical method used to correct for multiple comparisons for independent or weakly dependent test statistics. Although FDR control is frequently applied to microarray data analysis, gene expression is usually correlated, which might lead to inaccurate estimates. In this paper, we evaluate the accuracy of FDR estimation.
Using two real data sets, we resampled subgroups of patients and recalculated statistics of interest to illustrate the imprecision of FDR estimation. Next, we generated many simulated data sets with block correlation structures and realistic noise parameters, using the Ultimate Microarray Prediction, Inference, and Reality Engine (UMPIRE) R package. We estimated FDR using a beta-uniform mixture (BUM) model, and examined the variation in FDR estimation.
The three major sources of variation in FDR estimation are the sample size, correlations among genes, and the true proportion of differentially expressed genes (DEGs). The sample size and proportion of DEGs affect both magnitude and precision of FDR estimation, while the correlation structure mainly affects the variation of the estimated parameters.
We have decomposed various factors that affect FDR estimation, and illustrated the direction and extent of the impact. We found that the proportion of DEGs has a significant impact on FDR; this factor might have been overlooked in previous studies and deserves more thought when controlling FDR.
Vascular endothelial growth factor-2 (VEGFR-2 or KDR) is a known endothelial target also expressed in NSCLC tumor cells. We investigated the association between alterations in the KDR gene and clinical outcome in patients with resected NSCLC (n=248). KDR copy number gains (CNGs), measured by quantitative PCR and fluorescence in situ hybridization, were detected in 32% of tumors and associated with significantly higher KDR protein and higher microvessel density than tumors without CNGs. KDR CNGs were also associated with significantly increased risk of death (HR=5.16; P=0.003) in patients receiving adjuvant platinum-based chemotherapy, but no differences were observed in patients not receiving adjuvant therapy. To investigate potential mechanisms for these associations we assessed NSCLC cell lines and found that KDR CNGs were significantly associated with in vitro resistance to platinum chemotherapy as well as increased levels of nuclear HIF-1α in both NSCLC tumor specimens and cell lines. Furthermore, KDR knockdown experiments using small interfering RNA reduced platinum resistance, cell migration, and HIF-1α levels in cells bearing KDR CNGs, providing evidence for direct involvement of KDR. No KDR mutations were detected in exons 7, 11 and 21 by PCR-based sequencing; however, two variant SNP genotypes were associated with favorable overall survival in adenocarcinoma patients. Our findings suggest that tumor cell KDR CNGs may promote a more malignant phenotype including increased chemoresistance, angiogenesis, and HIF-1α levels, and that KDR CNGs may be a useful biomarker for identifying patients at high risk for recurrence after adjuvant therapy, a group that may benefit from VEGFR-2 blockade.
Translation initiation and activity of eukaryotic initiation factor-alpha (eIF2α), the rate-limiting step of translation initiation, is often overactivated in malignant cells. Here, we investigated the regulation and role of eIF2α in acute promyelocytic (APL) and acute myeloid leukemia (AML) cells in response to all-trans retinoic acid (ATRA) and arsenic trioxide (ATO), the front-line therapies in APL. ATRA and ATO induce Ser-51 phosphorylation (inactivation) of eIF2α, through the induction of protein kinase C delta (PKCδ) and PKR, but not other eIF2α kinases, such as GCN2 and PERK in APL (NB4) and AML cells (HL60, U937, and THP-1). Inhibition of eIF2α reduced the expression of cellular proteins that are involved in apoptosis (DAP5/p97), cell cycle (p21Waf1/Cip1), differentiation (TG2) and induced those regulating proliferation (c-myc) and survival (p70S6K). PI3K/Akt/mTOR pathway is involved in regulation of eIF2α through PKCδ/PKR axis. PKCδ and p-eIF2α protein expression levels revealed a significant association between the reduced levels of PKCδ (P = 0.0378) and peIF2 (P = 0.0041) and relapses in AML patients (n = 47). In conclusion, our study provides the first evidence that PKCδ regulates/inhibits eIF2α through induction of PKR in AML cells and reveals a novel signaling mechanism regulating translation initiation.
The Forkhead transcription factors (FOXO) are tumor suppressor genes regulating differentiation, metabolism, and apoptosis that functionally interact with signal transduction pathways shown to be deregulated and prognostic in acute myelogenous leukemia (AML). This study evaluated the level of expression and the prognostic relevance of total and phosphorylated FOXO3A protein in AML.
We used reverse-phase protein array methods to measure the level of total and phosphoprotein expression of FOXO3A, in leukemia-enriched protein samples from 511 newly diagnosed AML patients.
The expression range was similar to normal CD34+ cells and similar in blood and marrow. Levels of total FOXO3A were higher at relapse compared with diagnosis. Levels of pFOXO3A or the ratio of phospho to total (PT) were not associated with karyotpe but were higher in patients with FLT3 mutations. Higher levels of pFOXO3A or PT-FOXO3A were associated with increased proliferation evidenced by strong correlation with higher WBC, percent marrow, and blood blasts and by correlation with higher levels of Cyclins B1, D1 and D3, pGSK3, pMTOR, and pStat5. Patients with High levels of pFOXO3A or PT-FOXO3A had higher rates of primary resistance and shorter remission durations, which combine to cause an inferior survival experience (P = 0.0002). This effect was independent of cytogenetics. PT-FOXO3A was a statistically significant independent predictor in multivariate analysis.
High levels of phosphorylation of FOXO3A is a therapeutically targetable, independent adverse prognostic factor in AML.
The identification of key pathways dysregulated in non-small cell lung cancer (NSCLC) is an important step toward understanding lung pathogenesis and developing new therapeutic approaches.
Toward this goal, reverse-phase protein lysate arrays (RPPA) were used to compare signaling pathways between NSCLC tumors and paired normal lung tissue from 46 patients and assess their association with clinical outcome.
After RPPA quantification of 63 proteins and phosphoproteins, tissue pairs were randomized to a training set (n = 25 pairs) and test set (n = 21 pairs). In the training set, 15 protein markers were differentially expressed between tumors and normal lung (p ≤ 0.01), including markers in the PI3K/AKT and p38 MAPK signaling pathways (e.g., p70S6K, S6, p38, and phospho-p38), as well as caveolin-1 and β-catenin. A four-protein signature (p70S6K, cyclin B1, pSrc(Y527), and caveolin-1) independent of histology classified specimens as tumor versus normal with a predicted accuracy of 83%, sensitivity of 67%, and specificity of 100%. The signature was validated in the test set, correctly classifying all normal tissues and 14 of 21 tumor tissues. RPPA results were confirmed by immunohistochemistry for caveolin-1 and p70S6K. In tumors from patients with resected NSCLC, expression of proteins in the energy-sensing AMPK pathway (pLKB1, AMPK, p-Acetyl-CoA, pTSC2), adhesion, EGFR, and Rb signaling pathways was inversely associated with NSCLC recurrence.
These data provide evidence for dysregulation of several pathways including those involving energy sensing and adhesion that are potentially associated with NSCLC pathogenesis and disease recurrence.
NSCLC; Proteomics; Recurrence; AMPK; Adhesion