Advertisement
Research Article

Genomic Profiling Identifies GATA6 as a Candidate Oncogene Amplified in Pancreatobiliary Cancer

  • Kevin A. Kwei equal contributor,

    equal contributor Contributed equally to this work with: Kevin A. Kwei, Murali D. Bashyam

    Affiliation: Department of Pathology, Stanford University, Stanford, California, United States of America

    X
  • Murali D. Bashyam equal contributor,

    equal contributor Contributed equally to this work with: Kevin A. Kwei, Murali D. Bashyam

    Affiliations: Department of Pathology, Stanford University, Stanford, California, United States of America, Laboratory of Molecular Oncology, Centre for DNA Fingerprinting and Diagnostics, Nacharam, Hyderabad, India, National Genomics and Transcriptomics Facility, Centre for DNA Fingerprinting and Diagnostics, Nacharam, Hyderabad, India

    These authors are joint senior authors on this work.

    X
  • Jessica Kao,

    Affiliation: Department of Pathology, Stanford University, Stanford, California, United States of America

    X
  • Raman Ratheesh,

    Affiliation: Laboratory of Molecular Oncology, Centre for DNA Fingerprinting and Diagnostics, Nacharam, Hyderabad, India

    X
  • Edumakanti C. Reddy,

    Affiliation: National Genomics and Transcriptomics Facility, Centre for DNA Fingerprinting and Diagnostics, Nacharam, Hyderabad, India

    X
  • Young H. Kim,

    Affiliation: Department of Pathology, Stanford University, Stanford, California, United States of America

    X
  • Kelli Montgomery,

    Affiliation: Department of Pathology, Stanford University, Stanford, California, United States of America

    X
  • Craig P. Giacomini,

    Affiliation: Department of Pathology, Stanford University, Stanford, California, United States of America

    X
  • Yoon-La Choi,

    Affiliations: Department of Pathology, Stanford University, Stanford, California, United States of America, Department of Pathology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, South Korea

    X
  • Sreejata Chatterjee,

    Affiliation: Laboratory of Molecular Oncology, Centre for DNA Fingerprinting and Diagnostics, Nacharam, Hyderabad, India

    X
  • Collins A. Karikari,

    Affiliation: Department of Pathology, The Johns Hopkins University, Baltimore, Maryland, United States of America

    X
  • Keyan Salari,

    Affiliations: Department of Pathology, Stanford University, Stanford, California, United States of America, Department of Genetics, Stanford University, Stanford, California, United States of America

    X
  • Pei Wang,

    Affiliation: Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, Washington, United States of America

    X
  • Tina Hernandez-Boussard,

    Affiliation: Department of Genetics, Stanford University, Stanford, California, United States of America

    X
  • Gowrishankar Swarnalata,

    Affiliation: Department of Pathology, Apollo Hospitals, Hyderabad, India

    X
  • Matt van de Rijn,

    Affiliation: Department of Pathology, Stanford University, Stanford, California, United States of America

    X
  • Anirban Maitra,

    Affiliation: Department of Pathology, The Johns Hopkins University, Baltimore, Maryland, United States of America

    X
  • Jonathan R. Pollack mail

    pollack1@stanford.edu

    Affiliation: Department of Pathology, Stanford University, Stanford, California, United States of America

    These authors are joint senior authors on this work.

    X
  • Published: May 23, 2008
  • DOI: 10.1371/journal.pgen.1000081

Abstract

Pancreatobiliary cancers have among the highest mortality rates of any cancer type. Discovering the full spectrum of molecular genetic alterations may suggest new avenues for therapy. To catalogue genomic alterations, we carried out array-based genomic profiling of 31 exocrine pancreatic cancers and 6 distal bile duct cancers, expanded as xenografts to enrich the tumor cell fraction. We identified numerous focal DNA amplifications and deletions, including in 19% of pancreatobiliary cases gain at cytoband 18q11.2, a locus uncommonly amplified in other tumor types. The smallest shared amplification at 18q11.2 included GATA6, a transcriptional regulator previously linked to normal pancreas development. When amplified, GATA6 was overexpressed at both the mRNA and protein levels, and strong immunostaining was observed in 25 of 54 (46%) primary pancreatic cancers compared to 0 of 33 normal pancreas specimens surveyed. GATA6 expression in xenografts was associated with specific microarray gene-expression patterns, enriched for GATA binding sites and mitochondrial oxidative phosphorylation activity. siRNA mediated knockdown of GATA6 in pancreatic cancer cell lines with amplification led to reduced cell proliferation, cell cycle progression, and colony formation. Our findings indicate that GATA6 amplification and overexpression contribute to the oncogenic phenotypes of pancreatic cancer cells, and identify GATA6 as a candidate lineage-specific oncogene in pancreatobiliary cancer, with implications for novel treatment strategies.

Author Summary

Pancreatic cancer is a devastating disease, having among the lowest survival rates of any cancer. A better understanding of the molecular basis of pancreatic cancer may lead to improved rationale therapies. We report here the discovery of amplification (i.e. extra copies) of the GATA6 gene in many human pancreatic cancers. GATA6 is a regulator of gene expression and functions in the development of the normal pancreas. Our findings indicate that its amplification and aberrant overexpression contribute to pancreatic cancer development. GATA6 joins a growing list of cancer genes with key roles in normal human development but pathogenic roles in cancer when aberrantly expressed. Our discovery of GATA6 amplification provides a new foothold into understanding the pathogenic mechanisms underlying pancreatic cancer, and suggests new strategies for therapy by targeting GATA6 or the genes it regulates.

Introduction

Pancreatic cancer has among the highest mortality rates of any cancer, pointing to a critical need for more effective therapies. While much progress has been made in understanding pancreatic cancer pathogenesis, a more comprehensive characterization of molecular genetic alterations is needed to define new molecular targets and therapeutic opportunities [1].

Genomic DNA copy number alterations (CNAs) are frequent in pancreatic cancer, where they alter the dosage and expression of cancer genes. Amplified oncogenes include KRAS (also commonly activated by point mutation), AKT2 and MYB. Likewise, deleted tumor suppressor genes (TSGs) include CDKN2A, TP53 and SMAD4 (also inactivated by mutation and promoter hypermethylation) [2]. Mapping CNAs has become an important starting point for discovering new cancer genes, and indeed led to the original identification of CDKN2A and SMAD4 as TSGs [3],[4].

During development, the ventral portion of the pancreas arises from the primitive bile duct [5]. While less is known of extrahepatic bile duct cancers, they appear to share many features with pancreatic cancers, including frequent molecular alterations of KRAS, CDKN2A, TP53 and SMAD4, as well as global patterns of allelic loss [6],[7]. Because of their anatomic proximity and similar histologies, pancreatic and distal bile duct cancers can at times be difficult to distinguish, and from a clinical and research standpoint are often practically combined under the umbrella of pancreatobiliary cancer.

Recently, array-based comparative genomic hybridization (array CGH) has provided a powerful approach to catalog CNAs in cancer genomes [8],[9]. Profiling pancreatobiliary cancers, however, presents unique technical challenges due to the strong stromal response, where tumor cells typically account for less than 20% of cells in the specimen. Not surprisingly, to date genomic profiling of pancreatic cancer has been largely confined to studies of derived cell lines [10][13]. Strategies to enrich for tumor cells from primary cancers include physical microdissection, which is both technically demanding and, with the low yield of genomic DNA, is subject to biases in subsequent target amplification. An alternative strategy is expanding tumors as xenografts in nude mice, which effectively enriches the tumor cell fraction to >95%, (with the remaining stromal cells being of murine origin), while preserving genomic alterations of the parental tumor [14]. Here, we apply genomic profiling to a set of pancreatic and distal bile duct cancers grown as xenografts in nude mice, where among other alterations we identify and characterize GATA6 as a novel candidate lineage-specific oncogene amplified in pancreatobiliary cancer.

Results

To comprehensively catalog CNAs in pancreatobiliary cancers, we carried out array CGH-based genomic profiling of a set of 37 cancers (31 exocrine pancreatic cancers and 6 distal bile duct cancers) expanded as xenografts to enrich for tumor cells, using cDNA microarrays representing ~22,000 genes with a median interprobe spacing of ~15 Kb. We identified numerous CNAs, among which 17 focal high-level DNA amplifications (i.e. fluorescence ratios ≥3, corresponding to at least 5-fold amplification [8]) and 7 presumptive homozygous deletions (i.e. fluorescence ratios ≤0.25) were particularly informative in pinpointing known or novel candidate cancer genes (Table 1). By profiling gene expression in parallel, we also defined the subset of amplified genes exhibiting elevated expression (Table 1), a characteristic of oncogenes. For a subset of presumptive homozygous deletions, we validated homozygous loss by polymerase chain reaction (PCR) using human gene-specific primers (Table 1 and Figure S1).

thumbnail

Table 1. High-level amplifications and homozygous deletions.

doi:10.1371/journal.pgen.1000081.t001

Among the focal amplifications, we identified gain at 18q11.2 in 19% of pancreatobiliary cancers (5 of 31 pancreatic, and 2 of 6 bile duct). Notably, we found gains spanning 18q11.2 to be less common in other tumor types we had profiled on the same array platform, including cancers of the breast (3 gains in 89 (3%) tumors, 1 in 49 (2%) cell lines), prostate (0 in 64 (0%) tumors), lung (4 in 76 (5%) tumors, 4 in 52 (8%) cell lines) and colon (1 in 29 (3%) cell lines) [15][17] (and unpublished data), and in these other tumor types the gains when present were not focal, suggesting the putative driver oncogene within this locus may be specific to pancreatobiliary cancer. Strikingly, the smallest shared region of amplification among the xenograft specimens spanned just two annotated genes, GATA6 (GATA binding protein 6) and CTAGE1 (cutaneous T-cell lymphoma (CTCL)-associated antigen 1) (Figure 1A). GATA6 [18] belongs to the GATA factor family of transcriptional regulators, whose members are expressed in distinct developmental and tissue-specific patterns and regulate cell-restricted programs of gene expression [19]. Because GATA6 was known to regulate normal pancreas development [20],[21], we sought to explore a possible functional connection of GATA6 gene amplification and pancreatobiliary cancer.

thumbnail

Figure 1. GATA6 is focally amplified in pancreatobiliary cancer.

(A) Genomic profiles by CGH on cDNA microarrays of pancreatic (P) and bile duct (B) cancer xenografts across cytoband 18q11.2. Genes are ordered by genome position. Red indicates positive tumor/normal aCGH ratios (scale shown), and samples called gained at 18q11.2 are marked below by closed circle (gains highlighted in yellow). Genes and ESTs (IMAGE clone ID shown) on the microarray residing within the amplicon core are indicated. CTAGE1 (asterisked) was not present on the array but resides where shown. (B) Genomic profile of B291 by CGH on an Agilent ultra high-definition custom microarray tiling 18q11.2, mapped onto the UCSC genome browser (http://genome.ucsc.edu) [55]. The amplicon peak spans two genes, GATA6 and CTAGE1. (C) Q-PCR validation of GATA6 amplification in B291. Note, hybridization measurements by CGH tend to underestimate true CNA ratios [8]. (D) FISH validation of GATA6 amplification in the parent tumor (paraffin section) from which xenograft B291 was derived (left), and GATA6 gain in pancreatic cancer cell lines AsPC1 (center) and Panc3.27 (right). Gain is evident by the increased ratio of GATA6(red)/centromere-18(green) signals. DAPI (nuclear) counterstaining is shown in grayscale.

doi:10.1371/journal.pgen.1000081.g001

A single bile duct cancer xenograft specimen (B291) with focal high-level DNA amplification was particularly informative in defining amplicon boundaries. Using a custom Agilent ultra-high definition CGH array with probes tiling 18q11.2 with an average 343nt spacing, we first confirmed the amplicon boundaries in B291, finding the amplicon peak indeed spanned just GATA6 and CTAGE1 (Figure 1B). We also validated GATA6 amplification in B291 by quantitative (Q)-PCR (Figure 1C), and by fluorescence in situ hybridization (FISH) in the parent tumor from which the B291 xenograft was derived (Figure 1D, left panel), the latter excluding the possibility of amplification arising during xenograft growth. Focal 18q11.2 gain was also present in 3 of 18 (17%) pancreatic cancer cell lines (AsPC1, Panc3.27 and Capan1) we had previously profiled by array CGH ([13]) (Figure 1A, and data not shown), a finding we confirmed by FISH (Figure 1D).

Consistent with an oncogenic role, GATA6 exhibited increased mRNA expression by microarray in B291 (Figure 2A) and among the group of xenograft specimens with 18q11.2 gain (Figure 2B; P = 0.003, Mann-Whitney U-Test), a finding also confirmed by Q-reverse transcription (RT)-PCR (Figure 2C). In contrast, expression of the neighboring gene CTAGE1 was not detectable by Q-RT-PCR (data not shown). We also observed increased GATA6 protein levels by Western blot in pancreatic cancer cell lines with 18q11.2 gain, compared to pancreatic cancer cell lines without gain or to the nontumorigenic human pancreatic ductal epithelial line HPDE (Figure 2D), and by immunohistochemistry (IHC) in the parent tumor from which the B291 xenograft was derived (Figure 2E). To assess the frequency with which GATA6 exhibited elevated expression in primary pancreatic cancer, we performed IHC on a tissue microarray (TMA) that included cases of normal pancreas, pancreatitis and pancreatic ductal adenocarcinoma. We observed moderate and strong GATA6 nuclear staining respectively in 15 (28%) and 25 (46%) of 54 primary pancreatic cancers compared to just 3 (9%) and 0 (0%) of 33 normal pancreas specimens surveyed (P<0.001, χ2 test) (Figure 3). GATA6 expression was also elevated in pancreatitis (Figure 3E). There was no significant relation between GATA6 staining and tumor grade (P = 0.18, χ2 test).

thumbnail

Figure 2. GATA6 is overexpressed when amplified.

(A) Plot of DNA (by array CGH) vs. mRNA (by expression profiling) ratios for genes on chromosome 18 for specimen B291 shows GATA6 (indicated) to be the most highly expressed gene within the 18q11.2 amplicon. (B) GATA6 mRNA levels, measured by microarray, are elevated in pancreatobiliary xenografts with compared to without DNA gain at 18q11.2 (GATA6). Box plots show 25th, 50th and 75th percentiles; P-values (Mann-Whitney U-Test) for pairwise comparisons are indicated. (C) Q-RT-PCR validation of microarray-measured GATA6 transcript levels in eight specimens, four each with and without 18q11.2 gain. (D) Western blot analysis of representative pancreatic cancer cell lines indicates GATA6 (56 kD) is overexpressed at the protein level when amplified; GAPDH serves as a loading control. (E) IHC analysis of GATA6 protein expression (nuclear brown staining) indicates elevated expression in the parent tumor from which xenograft B291 was derived (left), in comparison to normal pancreatic duct from the same paraffin section (right).

doi:10.1371/journal.pgen.1000081.g002
thumbnail

Figure 3. GATA6 is overexpressed in primary pancreatic tumors.

Shown are representative IHC stains for GATA6 protein expression in (A) normal pancreas, and in pancreatic ductal adenocarcinoma with (B) absent, (C) moderate, and (D) strong nuclear staining. Filled arrowheads indicate pancreatic ductal epithelial cells (A), or pancreatic adenocarcinoma cells (BD). Open arrowhead (A) shows non-specific cytoplasmic staining observed in pancreatic acinar cells. (E) Distribution of GATA6 expression among different diagnoses represented on the tissue microarray. The IHC staining score considers both staining intensity and fraction of cells with nuclear staining (see Materials and Methods). GATA6 expression is significantly elevated in pancreatic cancer compared to normal pancreas (P<0.001, χ2 test).

doi:10.1371/journal.pgen.1000081.g003

Since GATA6 is a transcriptional regulator, we sought to identify co-expressed genes, which might include its downstream transcriptional targets and suggest functional involvements. Using Significance Analysis of Microarrays (SAM) [22], we identified 86 genes whose expression was significantly (False discovery rate, FDR, <1%) increased (73 genes) or decreased (13 genes) in xenografts with elevated GATA6 expression (Figure 4A). The SAM-identified gene set spanned diverse biological processes, and included known cancer genes like FGF1 and EVI1. Gene Set Enrichment Analysis (GSEA) [23] confirmed an enrichment of putative upstream GATA factor binding sites among the genes whose expression correlated with elevated GATA6 levels (P = 0.004) (Figure 4B). Interestingly, by GSEA the top functional gene sets associated with elevated GATA6 expression all related to mitochondrial activities connected to oxidative phosphorylation (Figure 4C).

thumbnail

Figure 4. GATA6 expression signature.

(A) Heatmap representation of genes identified by SAM analysis with significantly (FDR<1%) increased (73 genes) or decreased (13 genes) expression in xenografts with GATA6 mRNA levels above the mean. Specimens are ordered by GATA6 expression level; genes are ordered in descending rank of their SAM score. Expression levels are indicated by colorimetric ratio-scale (shown). (B) GSEA identifies enrichment of genes with putative GATA binding sites in xenografts with GATA6 expression levels above the mean. Enrichment is evidenced by the early positive deflection of the Kolmogorov-Smirnov running sum. The significance of the maximum running sum (S) was evaluated by comparison to 500 trials with randomly permuted class labels; the P-value is the frequency that S in the actual data is equaled or exceeded in the permuted data. (C) Top ranking (FDR shown) functional gene sets identified by GSEA to be enriched in xenografts with above-average GATA6 expression levels. TCA: tricarboxylic acid (Krebs) cycle.

doi:10.1371/journal.pgen.1000081.g004

To directly assess the functional significance of GATA6 amplification and overexpression in pancreatic cancer, we used RNA interference (RNAi) to target GATA6 knockdown in two pancreatic cancer cell lines, AsPC1 and Panc3.27, with GATA6 gain and overexpression. Transfection of two independent On-TARGETplus short interfering RNAs (siRNAs) targeting GATA6, designed and chemically modified to minimize off-target effects [24],[25], led to decreased GATA6 protein levels (Figure 5A), and to decreased cell proliferation compared to a negative control siRNA pool (Figure 5B). While the reduction in cell proliferation was relatively modest, it was statistically significant and reproducible in multiple independent experiments (not shown). In contrast, siRNA transfection of a pancreatic cancer cell line, PL45, without GATA6 amplification and overexpression (Figure 2D) did not diminish cell proliferation (Figure 5B, right panel), supporting the specificity of GATA6 targeting. We examined in more detail the effect of GATA6 knockdown in AsPC1 cells, where the reduced cell proliferation was attributable to decreased cell-cycle progression (as evidenced by decreased S-phase fraction; Figure 5C) but not increased apoptosis (Figure 5D). GATA6 knockdown in AsPC1 cells also led to reduced colony formation in liquid culture (Figure 5E).

thumbnail

Figure 5. GATA6 amplification/overexpression contributes to cell proliferation.

(A) Confirmation of siRNA-mediated knockdown of GATA6 in AsPC1 and Panc3.27 cells. Two different siRNAs (GATA6-1 and GATA6-2) were used to target GATA6, along with a non-targeting siRNA pool (control). GATA6 levels assayed by Western blot; GAPDH levels provide a loading control. (B) GATA6 knockdown results in decreased cell proliferation in reduced serum, measured by WST-1 assay, in cells with (AsPC1, Panc3.27) but not without (PL45) GATA6 gain/overexpression. *, P<0.05; **, P<0.01 (Student's t-test; GATA6 compared to control). (C) GATA6 knockdown reduces cell-cycle progression in AsPC1 cells, evidenced by decreased S-phase fraction following BrdU labeling, quantified by flow cytometry. *, P<0.05; (Student's t-test; GATA6 compared to control). (D) GATA6 knockdown does not significantly alter levels of apoptosis, quantified by annexin V staining. (E) GATA6 knockdown reduces colony growth of AsPC1 cells in liquid culture. Box plot illustrates 25th, mean and 75th percentile; P values (Student's t-test) indicated. Representative fields of Giemsa-stained colonies are shown (right).

doi:10.1371/journal.pgen.1000081.g005

In complementary experiments, we attempted to overexpress GATA6 by retroviral transduction in nontumorigenic human pancreatic ductal epithelial HPDE cells, and in the pancreatic cancer cell line PL45 harboring activated KRAS but no 18q11.2 gain. Though GATA6 expression was initially detected in infected cells by Western blot (data not shown), expression was lost upon expansion of cell pools under selection, suggesting GATA6 conferred negative fitness in these cell contexts.

Discussion

A main objective of our study was to comprehensively catalog CNAs in pancreatobiliary cancers. Genomic profiling of 31 pancreatic and 6 distal bile duct cancer xenografts identified numerous focal high-level DNA amplifications and homozygous deletions, thereby pinpointing known and candidate cancer genes. Known cancer genes included focal amplifications of MYC (8q24.21), KRAS (12p12.1) and AKT2 (19q13.2), and homozygous deletions of TGFBR2 (3p24.1) and CDKN2A (9p21.3). Other focal changes suggest entirely new pathobiology. For example, homozygous deletion of TLR3 (Toll-like receptor 3) (4q35.1), which functions in the innate immune response and is also highly expressed in pancreas [26],[27], suggests a possible role of infection in pancreatic carcinogenesis.

Prominent among novel oncogene candidates we identified GATA6 amplification at 18q11.2. GATA6 is one of six members of the mammalian GATA family of transcriptional regulators, each having two zinc finger domains and binding the common DNA sequence element (A/T)GATA(A/G) [19]. GATA factors 1–3 are expressed mainly in hematopoietic lineages, while GATA factors 4–6 are expressed in various tissues derived from mesoderm and endoderm, including the heart, liver, lung, gut, ovary and testis, where they function in cell lineage specification [19],[28]. In relation to pancreas development in the mouse, GATA4 and GATA6 are expressed in both endocrine and exocrine cell precursors, while in the adult pancreas expression of GATA4 and GATA6 is restricted to the exocrine and endocrine compartment, respectively [20],[29]. Recently, GATA4 and GATA6 have both been shown to be required for normal pancreas specification and development [20],[21]. While GATA6 has been linked to pancreas development, our findings now also connect GATA6 to pancreatic cancer, where GATA6 amplification and resultant overexpression contribute significantly (albeit at modest levels) to oncogenic phenotypes (cell proliferation, cell-cycle progression and colony formation) of pancreatic cancer cells.

Given its connection to development and cell specification, an oncogenic role of GATA6 might seem surprising. Indeed, GATA6 has been characterized as a TSG in other cell contexts [30],[31], and inactivating mutations have been identified in human malignant astrocytomas [31]. Nonetheless, other cell lineage-specific transcription factors have been found amplified in cancers, including MITF in melanoma [32], AR in hormone-independent prostate cancer [33], ESR1 in breast cancer [34], and most recently NKX2-1 (TITF) in lung cancer [17], [35][37]. The altered expression of such transcriptional regulators, having normal roles in lineage proliferation or survival, might be needed for tumor survival and progression in some cellular and genetic contexts, indicating a state of “lineage-dependency” [38]. More generally, the deregulated expression of transcription factors with roles in normal development reflects the principle of “oncology recapitulating ontogeny” [39]. While we detected GATA6 amplification primarily in pancreatobiliary cancers, GATA6 expression is not restricted to the developing pancreas, and therefore it remains to be determined whether GATA6 might have an oncogenic role in other cell lineages.

Another characteristic of lineage-specific oncogenes is that their oncogenic activity appears to be highly cell and genetic context dependent. MITF expression is growth inhibitory in normal human melanocytes [40], but in the context of BRAF activation (along with TP53 and RB1 pathway inactivation) leads to growth factor and anchorage independent growth [32]. Likewise, TITF1 is growth inhibitory when expressed in immortalized human lung epithelial cells [35], but promotes cell proliferation and survival when amplified in lung cancers [17],[37]. Consistent with these findings, GATA6 expression imparted negative fitness in immortalized human pancreatic ductal epithelial cells (HPDE), and in a pancreatic cancer cell line (PL45) with KRAS activation but no 18q11.2 gain. Additional studies are needed to clarify the genetic context of GATA6 oncogenic function.

While GATA6 was amplified in 19% of xenograft specimens, it was highly expressed at the protein level in 46% of primary pancreatic tumors surveyed. This finding suggests that GATA6 expression is likely elevated by mechanisms other than gene amplification in a substantial subset of cases. We also noted increased GATA6 expression in pancreatitis, which is a known risk factor for developing pancreatic cancer [41], and suggests a possible mechanistic link. As noted above, GATA4 is also expressed during normal pancreas development, though unlike GATA6 its expression is retained in the adult exocrine pancreas. Of interest, we have also observed DNA gains spanning GATA4 at 8p23.1 in a subset of xenografts (not shown), though none having focal DNA amplification. Additional studies are needed to examine the function, if any, of GATA4 in pancreatobiliary cancer.

Our analysis of gene-expression patterns correlated with GATA6 transcript levels has revealed an intriguing association with genes functioning in mitochondrial oxidative phosphorylation. Such altered mitochondrial activity might contribute to carcinogenesis through changes in cell metabolism, reactive oxygen species (ROS) production, or (through altered mitochondrial membrane potential) mitochondrial-associated apoptotic pathways [42],[43]. In the latter regard, it is notable in that we did not observe an effect of GATA6 knockdown on apoptosis, but rather on cell-cycle progression. While the exact connection to mitochondrial oxidative phosphorylation remains to be elucidated, it is of interest that a similar “OxPhos” expression pattern was recently identified in a subset of diffuse large B-cell lymphomas [44].

Since GATA6 is expressed in normal adult tissues like the endocrine pancreas, the lung, liver and heart [18], it is itself unlikely to become a useful target for therapy. However, future investigations will more precisely define the transcriptional effectors and pathways through which GATA6 mediates its oncogenic function, some of which might become important molecular targets. In conclusion, our genomic profiling and functional studies define GATA6 as a candidate lineage-specific oncogene in pancreatobiliary cancers, a finding which should lead to new opportunities for therapeutic intervention.

Materials and Methods

Specimens

Pancreatic and distal bile duct cancer xenografts were generated as described [14] at the Johns Hopkins Hospital, with Institutional Review Board (IRB) and Institutional Animal Care and Use Committee approval. Briefly, a 1 mm3 piece of the primary tumor was soaked in Matrigel (Collaborative Biomedical Research), then implanted subcutaneously in a nu/nu mouse. Engrafted tumors were harvested when they reached 1–2 cm in diameter. Tumor cell enrichment was confirmed by H&E-stained frozen section. From adjacent pieces of the xenograft, DNA was isolated using the Qiagen DNeasy Tissue kit, and RNA using the Trizol (Invitrogen) method. Pancreatic cancer cell lines were obtained from the American Type Culture Collection, and the HPDE cell line [45] was kindly provided by Dr. Ming Tsao (University of Toronto).

Array CGH and Expression Profiling

cDNA microarrays were obtained from the Stanford Functional Genomics Facility and included 39,632 human cDNAs, representing 22,279 mapped human genes (18,049 UniGene clusters [46], together with 4,230 additional mapped ESTs not assigned UniGene IDs). Array CGH and expression profiling were performed according to our published protocols [47],[48]. For array CGH, 4 µg of genomic DNA from each test sample was random-primer labeled with Cy5 and co-hybridized to the microarray along with 4 µg of Cy3-labeled sex-matched normal leukocyte reference DNA from a single donor. For gene-expression profiling, 50 µg of total RNA from each sample and 50 µg of “universal” reference RNA (derived from 11 different established human cell lines) were differentially labeled with Cy5 and Cy3, respectively, and co-hybridized to cDNA microarrays. Following overnight hybridization and washing, arrays were imaged using a GenePix 4000B scanner (Molecular Devices). Fluorescence ratios were extracted using SpotReader software (Niles Scientific), and the data uploaded into the Stanford Microarray Database (SMD) [49] for storage, retrieval and analysis. The complete microarray datasets are available at SMD and at the Gene Expression Omnibus (GEO) (accession GSE11152).

Microarray Data Analysis

Background-subtracted fluorescence ratios were normalized by mean centering genes for each array. For array CGH analysis, we included for subsequent analysis only well-measured genes with Cy3 reference-channel fluorescence signal intensity at least 1.4-fold above background in at least 50% of samples. Map positions for arrayed cDNA clones were assigned using the NCBI genome assembly, accessed through the UCSC genome browser database (NCBI Build 36). For genes represented by multiple arrayed cDNAs, the average fluorescence ratio was used. DNA gains and losses were identified by the fused lasso method [50]. We defined high-level DNA amplifications and presumptive homozygous deletions as contiguous regions identified by fused lasso with at least 50% of genes displaying fluorescence ratios ≥3 or ≤0.25, respectively. For expression profiling, fluorescence ratios were normalized for each array, and then well-measured genes (fluorescence intensities for the Cy5 or Cy3 channel at least 1.5-fold above background) were subsequently “mean-centered” (i.e. reported for each gene relative to the mean ratio across all samples).

SAM analysis [22] was performed using the 2-class method, comparing xenograft specimens with above and below mean GATA6 mRNA levels. GSEA [23] was carried out as described [51]. Genes with putative GATA binding sites (within the first 1-Kb upstream promoter sequence) were defined using MATCH software ([52]; default settings set to minimize false positives), applied to the common binding site matrix V$GATA_Q6 (all six GATA factors share a common DNA binding site, (A/T)GATA(A/G) [19]). To assess enrichment of GATA binding sites, the absolute value of the GSEA metric (Pearson correlation) was used in order to consider both upregulated and downregulated targets. GSEA using 522 functional gene sets was carried out as described [23].

High-Definition 18q11.12 CGH Microarray

A custom ultra high-definition CGH microarray was designed and obtained from Agilent Technologies [53]. The array included 4,362 probes tiling 1.5 Mb (Mb 17.5–19.0) of 18q11.2 with an average inter-probe spacing of 344 nt, and an additional 23,652 probes spanning the remaining genome for data normalization. DNAs were labeled as above, then hybridized to the Agilent array following the manufacturer's instructions, except using a 40 hr hybridization time (rather than the recommended 24 hr). Arrays were scanned using an Agilent G2505B scanner, and data extracted and normalized using Agilent Feature Extraction software (version 9.1) with default settings.

PCR Assays

To validate homozygous deletions (see Table 1), we used gene-specific primer-pairs to PCR amplify genomic DNA from xenograft specimens. Primer-pairs for genes flanking the regions of homozygous deletion, and designed to have a distinguishable fragment size, were included in the PCR reactions as internal controls, and normal DNA was used as positive control for primer pairs. PCR was performed on an Applied Biosystems GeneAmp 9700, using 40 ng DNA template, 1× PCR buffer (Applied Biosystems), 200 µM dNTPs, 2.0 mM MgCl2, 10 pmol each individual primer (Table S1), and 1 U of AmpliTaq Gold DNA polymerase (Applied Biosystems) in a 10 µl reaction. The reaction conditions were: 95°C 10 min initial denaturation, followed by 35 cycles (94°C 30 s; annealing temp (see Table S1) 30 s; 72°C 30 s), and a final extension of 72°C 7 min. PCR products were resolved by gel electrophoresis on a 1.8% TAE agarose gel, and visualized using a UVP gel documentation system.

To validate GATA6 amplification in B291, we carried out Q-PCR using the Quantitect SYBR Green PCR kit (Qiagen) on an ABI 7500 sequence detection system as per manufacturer's instructions. PCR was initiated at 95°C for 15 min (to activate the modified Taq polymerase), followed by a 40 cycle amplification (95°C 15 s, 58°C 30 s, 72°C 30 s). Melting curve analysis was performed to ensure specific PCR product while excluding primer dimers. We used the comparative CT method [54] to calculate relative DNA levels normalized to NPC1 (a gene located outside the 18q11.2 amplicon and not exhibiting CNA), which we then expressed as a ratio to the Ct value of GATA6 (also normalized to NPC1) obtained from normal DNA. PCR primer sequences are listed in Table S1.

To validate microarray-measured GATA6 mRNA levels, we carried out Q-RT-PCR using the QuantiTect SYBR Green RT-PCR Kit according to the manufacturer's instructions. The reaction mixture was first incubated at 50°C for 30 min for reverse transcription, then Q-PCR was carried out as above. For each specimen, relative levels of GATA6 transcript were calculated as the ratio of Ct value of GATA6 to that of GAPDH. Ratio values were then converted to log2 scale and normalized to the mean across all specimens. PCR primer sequences are listed in Table S1.

FISH

Probe labeling and FISH were performed using Vysis reagents according to the manufacturer's protocols. A locus-specific BAC mapping to GATA6 at 18q11.2 (RP11-1083G24; BACPAC Resources Centre) was labeled with SpectrumOrange, and co-hybridized with SpectrumGreen-labeled chromosome 18 centromere probe (CEP18; Vysis). Chromosomal locations of BACs were validated using normal metaphase slides (not shown). Slides were counterstained with DAPI, and imaged using an Olympus BX51 fluorescence microscope with Applied Imaging Cytovision 3.0 software.

Immunohistochemistry

A tissue microarray (TMA) was constructed using a tissue arrayer (Beecher Instruments) and archived formalin-fixed, paraffin-embedded pancreatic tissue specimens from Stanford University, with IRB approval. The TMA contained 1.2 mm cores representing normal pancreas (33 cases), pancreatitis (16), other benign diagnoses (9), and pancreatic ductal adenocarcinoma (54). For immunohistochemistry, a 4 µm section was cut from the tissue microarray block, de-paraffinized in Citrisolv (Fisher Scientific), and hydrated in a graded series of alcohol solutions. Heat-induced antigen retrieval was performed by microwave pretreatment in citrate (1 mM, pH 6.0) for 15 minutes before staining. Endogenous peroxidase was blocked by preincubation with 1% hydrogen peroxide in phosphate-buffered saline. A GATA6 mouse monoclonal antibody (R&D systems) was used at 1:10 dilution for 30 min. Chromogenic detection was carried out using a peroxidase-conjugated secondary antibody and DAB reagents provided with the Envision detection kit (Dako). Nuclear staining intensity (absent, weak, medium, strong; 0–3 scale) and fractional epithelium staining (up to 25%, 50%, 75%, 100%; 1–4 scale) were each recorded, then summed for a final staining score.

siRNA Transfections

On-TARGETplus siRNAs targeting GATA6, along with a negative control siRNA pool (ON-TARGETplus siCONTROL Non-targeting Pool), were obtained from Dharmacon. Sequences of siRNAs are listed in Table S1. Cell lines were maintained at 37°C in complete media of RPMI-1640 (Invitrogen), 10% FBS, 50 U/ml penicillin, and 50 U/ml streptomycin. For transfection, 100,000–250,000 cells were seeded per 6-well plate well, and transfected using Lipofectamine 2000 reagent (Invitrogen) according to the manufacturer's protocol. Cells were transfected with a final concentration of 50 nM siRNA for 6 hrs, subsequently replaced with reduced serum (2% FBS) growth media, where siRNA-mediated phenotypic effects were more reproducibly observed.

Western Blot

72 hours post-transfection, cells were lysed in 1× RIPA Lysis buffer (Upstate/Chemicon) supplemented with 1× Complete Protease Inhibitor (Roche, Indianapolis, IN), 0.1 mM sodium orthovanadate, 1 mM sodium fluoride and 1 mM PMSF, and protein quantified using the DC Protein Assay (Biorad). For Western blot, 25–40 µg protein lysate was electrophoresed on a 10% Tris/glycine polyacrylamide gradient gel (Biorad) and transferred to PVDF membrane (Biorad). After blocking in TBS-T buffer (20 mM Tris-HCl pH 7.4, 0.15 M NaCl, 0.1% Tween 20) with 5% dry milk for 30 min, blots were incubated sequentially with primary antibody at 4°C overnight, then HRP-conjugated secondary antibody at room temp for 45 min. Antibodies were used as follows: anti-GATA6 rabbit polyclonal antibody (1:200; Santa Cruz Biotechnology); anti-GAPDH rabbit polyclonal antibody (1:5,000 for loading control; Santa Cruz Biotechnology); HRP-conjugated anti-rabbit IgG (1:20,000, Pierce). Detection was carried out using the ECL kit (Amersham Biosciences).

Cell Proliferation Assay

24, 72 and 96 hours post-transfection, cell proliferation was quantified by colorimetry based on the metabolic cleavage of the tetrazolium salt WST-1 in viable cells, according to the manufacturer's protocol (Roche). WST-1 reagent was added at 1/10th the culture volume and incubated at 37°C for 30 min. Absorbance was then measured at 450 nm with reference to 650 nm using a Spectra Max 190 plate reader (Molecular Devices). Transfections were performed in triplicate and average (±1 SD) OD reported.

Cell-Cycle Analysis

72 hours post-transfection, cell-cycle distribution analysis was performed by flow cytometry using the BrdU-FITC Flow kit (BD Biosciences) per the manufacturer's instructions. Cells were incubated with 10 µM BrdU at 37°C for 4 hrs, then fixed and permeabilized with Cytofix/Cytoperm buffer (BD Biosciences). Cellular DNA was treated with DNase at 37° C for 1hr to expose incorporated BrdU, then cells were stained with anti-BrdU FITC antibody (to quantify incorporated BrdU) and 7-aminoactinomycin D (7-AAD; to quantify total DNA content). 10,000 events were scored by FACSCalibur (BD Biosciences) and analyzed using CellQuest software (BD Biosciences). Transfections were performed in triplicate and average (±1 SD) cell-cycle fractions reported.

Liquid Colony Formation Assay

24 hours post-transfection, 200 cells were plated each onto ten 10 cm dishes in complete media. After 2 weeks, surviving cells were stained with Giemsa (Sigma-Aldrich) for 15 min, and visible colonies counted on a light box.

Apoptosis Assay

72 hours post transfection, apoptosis was assayed by annexin V staining, quantified by flow cytometry using the Vybrant Apoptosis Assay kit (Invitrogen) as per the manufacturer's instructions. Floating cells and trypsinized adherent cells were pooled and resuspended in 200 µl annexin binding buffer. 2.5 µl Alexa Fluor 488 annexin V and 1 µl of 100 µg/ml propidium iodide (PI) solution were added and cells incubated for 15 min at room temp. Cells were then resuspended in equal volume of annexin binding buffer and analyzed immediately by flow cytometry. 10,000 events were scored by FACSCalibur and analyzed using CellQuest software. Transfections were performed in triplicate, and average (±1 SD) percent apoptosis reported.

Supporting Information

Figure S1.

Representative PCR-validations of homozygous deletion. (A) 3p24.1 deletion. TGFBR2, located within the presumptive homozygous deletion, is PCR-amplified from normal genomic DNA, but not from pancreatic cancer xenograft P224. (B) 9p21.2 deletion. MOBKL2B, within the deletion, is PCR-amplified from normal genomic DNA but not from P201.

doi:10.1371/journal.pgen.1000081.s001

(0.22 MB PDF)

Table S1.

Primer/siRNA sequences.

doi:10.1371/journal.pgen.1000081.s002

(0.02 MB XLS)

Acknowledgments

We wish to thank the SFGF for microarray manufacture, the SMD for database support, Ilana Galperin (Stanford Cytogenetics Laboratory) for assistance with FISH analysis, and Eon Rios for assistance with FACS analysis. We also thank the members of the Pollack lab for helpful discussions.

Author Contributions

Conceived and designed the experiments: KK MB AM JP. Performed the experiments: KK MB JK RR ER YK SC GS. Analyzed the data: KK MB JK RR ER YK CG YC SC KS PW GS JP. Contributed reagents/materials/analysis tools: KM CK PW TH MV AM. Wrote the paper: KK MB JP.

References

  1. 1. Bardeesy N, DePinho RA (2002) Pancreatic cancer biology and genetics. Nat Rev Cancer 2: 897–909.
  2. 2. Maitra A, Kern SE, Hruban RH (2006) Molecular pathogenesis of pancreatic cancer. Best Pract Res Clin Gastroenterol 20: 211–226.
  3. 3. Kamb A, Gruis NA, Weaver-Feldhaus J, Liu Q, Harshman K, et al. (1994) A cell cycle regulator potentially involved in genesis of many tumor types. Science 264: 436–440.
  4. 4. Hahn SA, Schutte M, Hoque AT, Moskaluk CA, da Costa LT, et al. (1996) DPC4, a candidate tumor suppressor gene at human chromosome 18q21.1. Science 271: 350–353.
  5. 5. Kozu T, Suda K, Toki F (1995) Pancreatic development and anatomical variation. Gastrointest Endosc Clin N Am 5: 1–30.
  6. 6. Rashid A (2002) Cellular and molecular biology of biliary tract cancers. Surg Oncol Clin N Am 11: 995–1009.
  7. 7. Iacobuzio-Donahue CA, van der Heijden MS, Baumgartner MR, Troup WJ, Romm JM, et al. (2004) Large-scale allelotype of pancreaticobiliary carcinoma provides quantitative estimates of genome-wide allelic loss. Cancer Res 64: 871–875.
  8. 8. Pollack JR, Perou CM, Alizadeh AA, Eisen MB, Pergamenschikov A, et al. (1999) Genome-wide analysis of DNA copy-number changes using cDNA microarrays. Nat Genet 23: 41–46.
  9. 9. Pinkel D, Albertson DG (2005) Array comparative genomic hybridization and its applications in cancer. Nat Genet 37: SupplS11–17.
  10. 10. Heidenblad M, Schoenmakers EF, Jonson T, Gorunova L, Veltman JA, et al. (2004) Genome-wide array-based comparative genomic hybridization reveals multiple amplification targets and novel homozygous deletions in pancreatic carcinoma cell lines. Cancer Res 64: 3052–3059.
  11. 11. Aguirre AJ, Brennan C, Bailey G, Sinha R, Feng B, et al. (2004) High-resolution characterization of the pancreatic adenocarcinoma genome. Proc Natl Acad Sci U S A 101: 9067–9072.
  12. 12. Holzmann K, Kohlhammer H, Schwaenen C, Wessendorf S, Kestler HA, et al. (2004) Genomic DNA-chip hybridization reveals a higher incidence of genomic amplifications in pancreatic cancer than conventional comparative genomic hybridization and leads to the identification of novel candidate genes. Cancer Res 64: 4428–4433.
  13. 13. Bashyam MD, Bair R, Kim YH, Wang P, Hernandez-Boussard T, et al. (2005) Array-based comparative genomic hybridization identifies localized DNA amplifications and homozygous deletions in pancreatic cancer. Neoplasia 7: 556–562.
  14. 14. Hahn SA, Seymour AB, Hoque AT, Schutte M, da Costa LT, et al. (1995) Allelotype of pancreatic adenocarcinoma using xenograft enrichment. Cancer Res 55: 4670–4675.
  15. 15. Bergamaschi A, Kim YH, Wang P, Sorlie T, Hernandez-Boussard T, et al. (2006) Distinct patterns of DNA copy number alteration are associated with different clinicopathological features and gene-expression subtypes of breast cancer. Genes Chromosomes Cancer 45: 1033–1040.
  16. 16. Lapointe J, Li C, Giacomini CP, Salari K, Huang S, et al. (2007) Genomic profiling reveals alternative genetic pathways of prostate tumorigenesis. Cancer Res 67: 8504–8510.
  17. 17. Kwei KA, Kim YH, Girard L, Kao J, Pacyna-Gengelbach M, et al. (2008) Genomic profiling identifies TITF1 as a lineage-specific oncogene amplified in lung cancer. Oncogene.
  18. 18. Suzuki H, Gabrielson E, Chen W, Anbazhagan R, van Engeland M, et al. (2002) A genomic screen for genes upregulated by demethylation and histone deacetylase inhibition in human colorectal cancer. Nat Genet 31: 141–149.
  19. 19. Molkentin JD (2000) The zinc finger-containing transcription factors GATA-4, -5, and -6. Ubiquitously expressed regulators of tissue-specific gene expression. J Biol Chem 275: 38949–38952.
  20. 20. Decker K, Goldman DC, Grasch CL, Sussel L (2006) Gata6 is an important regulator of mouse pancreas development. Dev Biol 298: 415–429.
  21. 21. Watt AJ, Zhao R, Li J, Duncan SA (2007) Development of the mammalian liver and ventral pancreas is dependent on GATA4. BMC Dev Biol 7: 37.
  22. 22. Tusher VG, Tibshirani R, Chu G (2001) Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci U S A 98: 5116–5121.
  23. 23. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, et al. (2005) Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 102: 15545–15550.
  24. 24. Birmingham A, Anderson EM, Reynolds A, Ilsley-Tyree D, Leake D, et al. (2006) 3′ UTR seed matches, but not overall identity, are associated with RNAi off-targets. Nat Methods 3: 199–204.
  25. 25. Jackson AL, Burchard J, Leake D, Reynolds A, Schelter J, et al. (2006) Position-specific chemical modification of siRNAs reduces “off-target” transcript silencing. Rna 12: 1197–1205.
  26. 26. Rock FL, Hardiman G, Timans JC, Kastelein RA, Bazan JF (1998) A family of human receptors structurally related to Drosophila Toll. Proc Natl Acad Sci U S A 95: 588–593.
  27. 27. Alexopoulou L, Holt AC, Medzhitov R, Flavell RA (2001) Recognition of double-stranded RNA and activation of NF-kappaB by Toll-like receptor 3. Nature 413: 732–738.
  28. 28. Patient RK, McGhee JD (2002) The GATA family (vertebrates and invertebrates). Curr Opin Genet Dev 12: 416–422.
  29. 29. Ketola I, Otonkoski T, Pulkkinen MA, Niemi H, Palgi J, et al. (2004) Transcription factor GATA-6 is expressed in the endocrine and GATA-4 in the exocrine pancreas. Mol Cell Endocrinol 226: 51–57.
  30. 30. Capo-chichi CD, Roland IH, Vanderveer L, Bao R, Yamagata T, et al. (2003) Anomalous expression of epithelial differentiation-determining GATA factors in ovarian tumorigenesis. Cancer Res 63: 4967–4977.
  31. 31. Kamnasaran D, Qian B, Hawkins C, Stanford WL, Guha A (2007) GATA6 is an astrocytoma tumor suppressor gene identified by gene trapping of mouse glioma model. Proc Natl Acad Sci U S A 104: 8053–8058.
  32. 32. Garraway LA, Widlund HR, Rubin MA, Getz G, Berger AJ, et al. (2005) Integrative genomic analyses identify MITF as a lineage survival oncogene amplified in malignant melanoma. Nature 436: 117–122.
  33. 33. Visakorpi T, Hyytinen E, Koivisto P, Tanner M, Keinanen R, et al. (1995) In vivo amplification of the androgen receptor gene and progression of human prostate cancer. Nat Genet 9: 401–406.
  34. 34. Holst F, Stahl PR, Ruiz C, Hellwinkel O, Jehan Z, et al. (2007) Estrogen receptor alpha (ESR1) gene amplification is frequent in breast cancer. Nat Genet 39: 655–660.
  35. 35. Tanaka H, Yanagisawa K, Shinjo K, Taguchi A, Maeno K, et al. (2007) Lineage-specific dependency of lung adenocarcinomas on the lung development regulator TTF-1. Cancer Res 67: 6007–6011.
  36. 36. Kendall J, Liu Q, Bakleh A, Krasnitz A, Nguyen KC, et al. (2007) Oncogenic cooperation and coamplification of developmental transcription factor genes in lung cancer. Proc Natl Acad Sci U S A 104: 16663–16668.
  37. 37. Weir BA, Woo MS, Getz G, Perner S, Ding L, et al. (2007) Characterizing the cancer genome in lung adenocarcinoma. Nature 450: 893–898.
  38. 38. Garraway LA, Sellers WR (2006) Lineage dependency and lineage-survival oncogenes in human cancer. Nat Rev Cancer 6: 593–602.
  39. 39. Lechner JF, Fugaro JM, Wong Y, Pass HI, Harris CC, et al. (2001) Perspective: cell differentiation theory may advance early detection of and therapy for lung cancer. Radiat Res 155: 235–238.
  40. 40. Loercher AE, Tank EM, Delston RB, Harbour JW (2005) MITF links differentiation with cell cycle arrest in melanocytes by transcriptional activation of INK4A. J Cell Biol 168: 35–40.
  41. 41. Algul H, Treiber M, Lesina M, Schmid RM (2007) Mechanisms of disease: chronic inflammation and cancer in the pancreas–a potential role for pancreatic stellate cells? Nat Clin Pract Gastroenterol Hepatol 4: 454–462.
  42. 42. Birch-Machin MA (2006) The role of mitochondria in ageing and carcinogenesis. Clin Exp Dermatol 31: 548–552.
  43. 43. Galluzzi L, Larochette N, Zamzami N, Kroemer G (2006) Mitochondria as therapeutic targets for cancer chemotherapy. Oncogene 25: 4812–4830.
  44. 44. Monti S, Savage KJ, Kutok JL, Feuerhake F, Kurtin P, et al. (2005) Molecular profiling of diffuse large B-cell lymphoma identifies robust subtypes including one characterized by host inflammatory response. Blood 105: 1851–1861.
  45. 45. Liu N, Furukawa T, Kobari M, Tsao MS (1998) Comparative phenotypic studies of duct epithelial cell lines derived from normal human pancreas and pancreatic carcinoma. Am J Pathol 153: 263–269.
  46. 46. Schuler GD (1997) Pieces of the puzzle: expressed sequence tags and the catalog of human genes. J Mol Med 75: 694–698.
  47. 47. Pollack JR, Sorlie T, Perou CM, Rees CA, Jeffrey SS, et al. (2002) Microarray analysis reveals a major direct role of DNA copy number alteration in the transcriptional program of human breast tumors. Proc Natl Acad Sci U S A 99: 12963–12968.
  48. 48. Lapointe J, Li C, Higgins JP, van de Rijn M, Bair E, et al. (2004) Gene expression profiling identifies clinically relevant subtypes of prostate cancer. Proc Natl Acad Sci U S A 101: 811–816.
  49. 49. Demeter J, Beauheim C, Gollub J, Hernandez-Boussard T, Jin H, et al. (2007) The Stanford Microarray Database: implementation of new analysis tools and open source release of software. Nucleic Acids Res 35: D766–770.
  50. 50. Tibshirani R, Wang P (2008) Spatial smoothing and hot spot detection for CGH data using the fused lasso. Biostatistics 9: 18–29.
  51. 51. Kim YH, Girard L, Giacomini CP, Wang P, Hernandez-Boussard T, et al. (2006) Combined microarray analysis of small cell lung cancer reveals altered apoptotic balance and distinct expression signatures of MYC family gene amplification. Oncogene 25: 130–138.
  52. 52. Kel AE, Gossling E, Reuter I, Cheremushkin E, Kel-Margoulis OV, et al. (2003) MATCH: A tool for searching transcription factor binding sites in DNA sequences. Nucleic Acids Res 31: 3576–3579.
  53. 53. Barrett MT, Scheffer A, Ben-Dor A, Sampas N, Lipson D, et al. (2004) Comparative genomic hybridization using oligonucleotide microarrays and total genomic DNA. Proc Natl Acad Sci U S A 101: 17765–17770.
  54. 54. Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2(−Delta Delta C(T)) Method. Methods 25: 402–408.
  55. 55. Kent WJ, Sugnet CW, Furey TS, Roskin KM, Pringle TH, et al. (2002) The human genome browser at UCSC. Genome Res 12: 996–1006.