Whole-genome Profiling Research Articles

Growth Phase-Dependent Activation of the DccRS Regulon of Campylobacter jejuni

Two-component systems are widespread prokaryotic signal transduction devices which allow the regulation of cellular functions in response to changing environmental conditions. The two-component system DccRS (Cj1223c-Cj1222c) of Campylobacter jejuni is important for the colonization of chickens. Here, we dissect the DccRS system in more detail and provide evidence that the sensor DccS selectively phosphorylates the cognate effector, DccR. Microarray expression profiling, real-time reverse transcription-PCR (RT-PCR), electrophoretic mobility shift assay, and primer extension analyses revealed that the DccRS regulon of strain 81116 consists of five promoter elements, all containing the consensus direct repeat sequence WTTCAC-N6-TTCACW covering the putative -35 promoter regions. One of these promoters is located in front of an operon encoding a putative macrolide efflux pump while the others are in front of genes coding for putative periplasmic or membrane proteins. The DccRS-regulated genes in C. jejuni strain 81116 are needed to enhance early in vivo growth of C. jejuni in 7-day-old chickens. The DccRS system is activated in the late stationary bacterial growth phase, probably by released metabolic products. Whole-genome mRNA profiling and real-time RT-PCR analysis under these conditions demonstrated that the system has no influence on the transcription of genes outside the DccRS regulon.

Genetic analysis of variation in transcription factor binding in yeast

Variation in transcriptional regulation is thought to be a major cause of phenotypic diversity1,2. Although widespread differences in gene expression among individuals of a species have been observed3-8, few studies have examined the variability of transcription factor (TF) binding, and thus the extent and underlying genetic basis of TF binding diversity is largely unknown. In this study, we mapped differences in transcription binding among individuals and elucidated the genetic basis of such variation on a genome-wide scale. Whole-genome Ste12 binding profiles were determined using ChIP Seq in pheromone-treated cells of 43 segregants of a cross between two highly diverged yeast strains and their parental lines. We identified extensive Ste12 binding variation among individuals and mapped underlying cis- and trans- acting loci responsible for such variation. We showed that the majority of TF binding variation is cis-linked and that many variations are associated with polymorphisms residing in the binding motifs of Ste12 as well as those of several proposed Ste12 cofactors. We also identified two trans factors, AMN1 and FLO8, that modulate Ste12 binding to promoters of more than 10 genes under α-factor treatment. Neither of these two genes was known to regulate Ste12 previously, and we suggest that they may be mediators of gene activity and phenotypic diversity. Ste12 binding strongly correlates with gene expression for more than 200 genes indicating that binding variation is functional. Many of the variable bound genes are involved in cell wall organization and biogenesis. Overall these studies identified genetic regulators of molecular diversity among individuals and provide novel insights into mechanisms of gene regulation.

Genome-wide analysis of aberrant methylation in human breast cancer cells using methyl-DNA immunoprecipitation combined with high-throughput sequencing

BackgroundCancer cells undergo massive alterations to their DNA methylation patterns that result in aberrant gene expression and malignant phenotypes. However, the mechanisms that underlie methylome changes are not well understood nor is the genomic distribution of DNA methylation changes well characterized.ResultsHere, we performed methylated DNA immunoprecipitation combined with high-throughput sequencing (MeDIP-seq) to obtain whole-genome DNA methylation profiles for eight human breast cancer cell (BCC) lines and for normal human mammary epithelial cells (HMEC). The MeDIP-seq analysis generated non-biased DNA methylation maps by covering almost the entire genome with sufficient depth and resolution. The most prominent feature of the BCC lines compared to HMEC was a massively reduced methylation level particularly in CpG-poor regions. While hypomethylation did not appear to be associated with particular genomic features, hypermethylation preferentially occurred at CpG-rich gene-related regions independently of the distance from transcription start sites. We also investigated methylome alterations during epithelial-to-mesenchymal transition (EMT) in MCF7 cells. EMT induction was associated with specific alterations to the methylation patterns of gene-related CpG-rich regions, although overall methylation levels were not significantly altered. Moreover, approximately 40% of the epithelial cell-specific methylation patterns in gene-related regions were altered to those typical of mesenchymal cells, suggesting a cell-type specific regulation of DNA methylation.ConclusionsThis study provides the most comprehensive analysis to date of the methylome of human mammary cell lines and has produced novel insights into the mechanisms of methylome alteration during tumorigenesis and the interdependence between DNA methylome alterations and morphological changes.

Identification of Stage-Specific Gene Modulation during Early Thymocyte Development by Whole-Genome Profiling Analysis after Aryl Hydrocarbon Receptor Activation

The aryl hydrocarbon receptor (AHR) is a basic helix-loop-helix transcription factor, implicated as an important modulator of the immune system and of early thymocyte development. We have shown previously that AHR activation by the environmental contaminant and potent AHR agonist 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) leads to a significant decline in the percentage of S-phase cells in the CD3(-)CD4(-)CD8(-) triple-negative stage (TN) 3 and TN4 T-cell committed thymocytes 9 to 12 h after exposure. In the more immature TN1- or TN2-stage cells, no effect on cell cycle was observed. To identify early molecular targets, which could provide insight into how the AHR acts as a modulator of thymocyte development and cell cycle regulation, we performed gene-profiling experiments using RNA isolated from four intrathymic progenitor populations in which the AHR was activated for 6 or 12 h. This microarray analysis of AHR activation identified 108 distinct gene probes that were significantly modulated in the TN1-4 thymocyte progenitor stages. Although most of the genes identified have specific AHR recognition sequences, only seven genes were altered exclusively in the two T-cell committed stages of early thymocyte development (TN3 and TN4) in which the decline of S-phase cells is seen. Moreover, all seven of these genes were reduced in expression, and five of the seven are associated with cell cycle regulatory processes. These seven genes are novel targets for modulation by the TCDD-activated AHR and may be involved in the observed cell-cycle arrest and suppression of early thymocyte development.

Mouse Prkar1a haploinsufficiency leads to an increase in tumors in the Trp53+/− or Rb1+/− backgrounds and chemically induced skin papillomas by dysregulation of the cell cycle and Wnt signaling

PRKAR1A inactivation leads to dysregulated cAMP signaling and Carney complex (CNC) in humans, a syndrome associated with skin, endocrine and other tumors. The CNC phenotype is not easily explained by the ubiquitous cAMP signaling defect; furthermore, Prkar1a(+/-) mice did not develop skin and other CNC tumors. To identify whether a Prkar1a defect is truly a generic but weak tumorigenic signal that depends on tissue-specific or other factors, we investigated Prkar1a(+/-) mice when bred within the Rb1(+/-) or Trp53(+/-) backgrounds, or treated with a two-step skin carcinogenesis protocol. Prkar1a(+/-) Trp53(+/-) mice developed more sarcomas than Trp53(+/-) mice (P < 0.05) and Prkar1a(+/-) Rb1(+/-) mice grew more (and larger) pituitary and thyroid tumors than Rb1(+/-) mice. All mice with double heterozygosity had significantly reduced life-spans compared with their single-heterozygous counterparts. Prkar1a(+/-) mice also developed more papillomas than wild-type animals. A whole-genome transcriptome profiling of tumors produced by all three models identified Wnt signaling as the main pathway activated by abnormal cAMP signaling, along with cell cycle abnormalities; all changes were confirmed by qRT-PCR array and immunohistochemistry. siRNA down-regulation of Ctnnb1, E2f1 or Cdk4 inhibited proliferation of human adrenal cells bearing a PRKAR1A-inactivating mutation and Prkar1a(+/-) mouse embryonic fibroblasts and arrested both cell lines at the G0/G1 phase of the cell cycle. In conclusion, Prkar1a haploinsufficiency is a relatively weak tumorigenic signal that can act synergistically with other tumor suppressor gene defects or chemicals to induce tumors, mostly through Wnt-signaling activation and cell cycle dysregulation, consistent with studies in human neoplasms carrying PRKAR1A defects.

The dissection of transcriptional modules regulated by various drugs of abuse in the mouse striatum

The transcriptional response to six commonly-abused drugs was assessed in the mouse brain revealing common modules of drug-induced genes.

Identifying Sleep Regulatory Genes Using aDrosophilaModel of Insomnia

Although it is widely accepted that sleep must serve an essential biological function, little is known about molecules that underlie sleep regulation. Given that insomnia is a common sleep disorder that disrupts the ability to initiate and maintain restorative sleep, a better understanding of its molecular underpinning may provide crucial insights into sleep regulatory processes. Thus, we created a line of flies using laboratory selection that share traits with human insomnia. After 60 generations, insomnia-like (ins-l) flies sleep 60 min a day, exhibit difficulty initiating sleep, difficulty maintaining sleep, and show evidence of daytime cognitive impairment. ins-l flies are also hyperactive and hyperresponsive to environmental perturbations. In addition, they have difficulty maintaining their balance, have elevated levels of dopamine, are short-lived, and show increased levels of triglycerides, cholesterol, and free fatty acids. Although their core molecular clock remains intact, ins-l flies lose their ability to sleep when placed into constant darkness. Whole-genome profiling identified genes that are modified in ins-l flies. Among those differentially expressed transcripts, genes involved in metabolism, neuronal activity, and sensory perception constituted over-represented categories. We demonstrate that two of these genes are upregulated in human subjects after acute sleep deprivation. Together, these data indicate that the ins-l flies are a useful tool that can be used to identify molecules important for sleep regulation and may provide insights into both the causes and long-term consequences of insomnia.

Uniparental disomies, homozygous deletions, amplifications, and target genes in mantle cell lymphoma revealed by integrative high-resolution whole-genome profiling

AbstractMantle cell lymphoma (MCL) is genetically characterized by the t(11;14)(q13;q32) translocation and a high number of secondary chromosomal alterations. However, only a limited number of target genes have been identified. We have studied 10 MCL cell lines and 28 primary tumors with a combination of a high-density single-nucleotide polymorphism array and gene expression profiling. We detected highly altered genomes in the majority of the samples with a high number of partial uniparental disomies (UPDs). The UPD at 17p was one of the most common, and it was associated with TP53 gene inactivation. Homozygous deletions targeted 4 known tumor suppressor genes (CDKN2C, BCL2L11, CDKN2A, and RB1) and 6 new genes (FAF1, MAP2, SP100, MOBKL2B, ZNF280A, and PRAME). Gene amplification coupled with overexpression was identified in 35 different regions. The most recurrent amplified regions were 11q13.3-q13.5, 13q31.3, and 18q21.33, which targeted CCND1, C13orf25, and BCL2, respectively. Interestingly, the breakpoints flanking all the genomic alterations, including UPDs, were significantly associated with genomic regions enriched in copy number variants and segmental duplications, suggesting that the recombination at these regions may play a role in the genomic instability of MCL. This integrative genomic analysis has revealed target genes that may be potentially relevant in MCL pathogenesis.

Whole-Genome Discovery and Profiling of Small RNAs Using the SOLiD™ System

Whole-Genome Discovery and Profiling of Small RNAs Using the SOLiD™ System

DETORQUEO, QUIRKY, and ZERZAUST Represent Novel Components Involved in Organ Development Mediated by the Receptor-Like Kinase STRUBBELIG in Arabidopsis thaliana

Intercellular signaling plays an important role in controlling cellular behavior in apical meristems and developing organs in plants. One prominent example in Arabidopsis is the regulation of floral organ shape, ovule integument morphogenesis, the cell division plane, and root hair patterning by the leucine-rich repeat receptor-like kinase STRUBBELIG (SUB). Interestingly, kinase activity of SUB is not essential for its in vivo function, indicating that SUB may be an atypical or inactive receptor-like kinase. Since little is known about signaling by atypical receptor-like kinases, we used forward genetics to identify genes that potentially function in SUB-dependent processes and found recessive mutations in three genes that result in a sub-like phenotype. Plants with a defect in DETORQEO (DOQ), QUIRKY (QKY), and ZERZAUST (ZET) show corresponding defects in outer integument development, floral organ shape, and stem twisting. The mutants also show sub-like cellular defects in the floral meristem and in root hair patterning. Thus, SUB, DOQ, QKY, and ZET define the STRUBBELIG-LIKE MUTANT (SLM) class of genes. Molecular cloning of QKY identified a putative transmembrane protein carrying four C2 domains, suggesting that QKY may function in membrane trafficking in a Ca2+-dependent fashion. Morphological analysis of single and all pair-wise double-mutant combinations indicated that SLM genes have overlapping, but also distinct, functions in plant organogenesis. This notion was supported by a systematic comparison of whole-genome transcript profiles during floral development, which molecularly defined common and distinct sets of affected processes in slm mutants. Further analysis indicated that many SLM-responsive genes have functions in cell wall biology, hormone signaling, and various stress responses. Taken together, our data suggest that DOQ, QKY, and ZET contribute to SUB-dependent organogenesis and shed light on the mechanisms, which are dependent on signaling through the atypical receptor-like kinase SUB.

602: Genome-wide array-based copy number profiling in placentas from unexplained stillbirth

While precise estimates vary, karyotypic abnormalities are present in 6-12% of stillbirths (SB). However, due to inherent limitations of the conventional cytogenetic techniques of FISH and karytoype, the contribution of genomic aberrations is likely underrepresented. We hypothesized that use of high-resolution genomic profiling by aCGH may overcome such limitations, and simultaneously allow for assessment of genomic copy number variants (CNV). We thus sought to determine by aCGH whether structural variants in placental tissue accompany SB in a prospectively attained, well-characterized case-control cohort. Prospectively acquired SB cases >22 weeks underwent amniocentesis for karyotype, maternal toxicology, serology and vaginal microbiology. At delivery multiple placental biopsies were obtained alongside fetal blood sampling, microbial sampling, complete placental histology and fetal autopsy. Classification of ”unexplained” SB was by strict criteria after multidisciplinary audit and included both appropriate (AGA) and small (<10th%) for gestational age (SGA). Placental DNA was extracted and tested by Illumina HumanCNV370-Duo chip to enable whole genomic CNV coverage. 24 SB meeting criteria for “unexplained SB” were studied by aCGH. These included 11 AGA and 13 SGA fetuses. Comparative analysis against AGA (n=10) and SGA (n=14) liveborns was performed. A total of 32 CNVs larger than 100 kb were detected in 18 placentas: 9/24 SB and 9/24 control placentas. Of the 32 detected CNVs, 24 were copy number gains and 8 were copy number losses. The detected CNVs were associated with different chromosomes and their sizes ranged from 142 kb to 2.9 Mb. Evaluation of stillbirths with aCGH using placental tissue is feasible, and enables whole-genome profiling of unexplained stillbirths. Our findings illustrate the power of such approaches to identify genomic variants, and lay the groundwork for future population-based analysis.

Molecular analysis of hormone receptor positive (luminal) breast cancers – What have we learnt?

Recently, whole-genome molecular profiling of cancers has revealed that breast cancer consists of a number of distinct diseases at the biological level, each of which will require independent research into the most suitable therapy for that patient. In particular, this has long confirmed the clinician’s impression that the clinical behaviour of oestrogen receptor (ER)-positive breast cancer can be markedly heterogeneous despite similar levels of expression of the oestrogen receptor. At present, it seems that there are at least two distinct diseases of luminal origins. In the future, it is likely that we will be treating the luminal-A tumours, characterised by high expression of the ER and related genes, differently from the non-luminal-A tumours, which are characterised by low expression of the ER and related genes, high expression of proliferation genes and a poor clinical outcome. This article reviews the progress thus far in producing a framework for defining the ER-positive luminal subtypes and for our current understanding of the genetic aberrations that may be contributing to the poor prognosis of the non-luminal-A breast cancers.

Abstract A7: Identification of novel genetic alterations in GI tumors using whole-genome profiling

Abstract A7 In our previous studies, we identified novel pancreatic cancer genes using a combination of array-based comparative genomic hybridization (aCGH) and gene expression microarrays (Neoplasia, 2007 and Kwei and PLOS Genet, 2008). In the current study, we have used a multipronged approach to address two important GI cancers prevalent in India. Colorectal cancer (CRC) is a major health problem worldwide and is usually an age-related disease. A small proportion of patients present at an early age due to familial cancer syndromes. Surprisingly, there appears to be a high proportion of patients in India who present at an early age without a family background and succumb to aggressive metastatic tumors. Using immunohistochemistry, mutation screening and microsatellite instability analysis of more than 150 samples, we show that Wnt signaling, a hallmark of CRC in the West, is not a major feature in a majority of young patients in India unlike older patients. There does not appear to be a significant difference between the two categories of CRC patients in other parameters including gender, lifestyle, diet, tumor location, frequency of microsatellite instability, and Loss of Heterozygosity at the BRCAI locus. In addition, we have identified several novel mutations in the APC gene in Wnt positive tumors. Determination of genome-wide copy number alterations and transcript profiles is expected to delineate the genetic aberrations that may lead to CRC in the young in India. Squamous cell carcinoma of the esophagus (ESCC) is common in India and Asia but rare in the West, as against esophagus adenocarcinoma which is common in the West but not in India. ESCC is a poorly studied highly aggressive cancer with survival rates below 15% in most populations. We have carried out molecular analysis of about 75 ESCC samples and 26 adenocarcinoma samples using archived as well as fresh resected tumors. Results indicate that unlike the ESCC samples, a significant proportion of adenocarcinoma samples exhibited an activated Wnt signaling pathway. Inactivation of p53 however appeared to be a common event in both Esophagus cancer subtypes. Analysis of copy number alterations using array-based CGH has revealed a recurrent amplification at 10q21 in ESCC samples. The amplification harbored genes that exhibited comparably altered transcript levels, as determined by array-based transcript profiling. We are also carrying out molecular analyses of mixed adeno-squamous tumors of the esophagus; results are expected to reveal whether the squamous and adeno components have an independent origin or vice versa. Citation Information: Cancer Prev Res 2008;1(7 Suppl):A7.

Pyramiding QTLs to improve malting quality in barley: gains in phenotype and genetic diversity

The ability of barley (Hordeum vulgare L.) breeders to deliver germplasm that combine elite malt quality characteristics, disease resistances, and important agronomic traits has been greatly enhanced by the use of molecular marker technologies. These technologies facilitate the rapid transfer of desirable traits from diverse, elite, germplasm into locally adapted varieties. This present study sought to obtain an additive genetic effect by combining favourable alleles associated with the malting quality of two elite donor parents (Harrington and Morex) such that the resultant progeny would possess quality superior to either parent. Analysis of genetic diversity, based on whole-genome profiling with 700 DArT markers, showed clear separation of the BC6F1-derived doubled haploid lines from existing malting barley germplasm, indicating they represent a distinctly different source population for genetic improvement. Micro-malting quality results of the BC-derived lines showed substantial quality improvements, compared with the recurrent parent. Malt extract levels were increased by 1.5–2.0%, while diastase levels increased from approximately 320 WKE to 400–460 WKE. Similarly, α-amylase levels were increased from 160 units to between 218 and 251 units, and wort viscosities lowered from 1.90 cps to 1.72–1.82 cps. Other quality improvements include increases in β-glucanase levels from 375 to between 447 and 512 units, and reductions in wort β-glucan levels by 30–60%. Whilst the genetic gains compared to the recurrent parent were impressive, quality of the derived lines were largely equivalent to the levels now available in the recently released varieties, Buloke and Flagship. In a few cases, the backcross-derived lines also showed similarities to the original donors, Harrington and Morex, but in none of the cases did quality of these lines exceed those of either Harrington or Morex.

Rapid whole-genome mutational profiling using next-generation sequencing technologies

Forward genetic mutational studies, adaptive evolution, and phenotypic screening are powerful tools for creating new variant organisms with desirable traits. However, mutations generated in the process cannot be easily identified with traditional genetic tools. We show that new high-throughput, massively parallel sequencing technologies can completely and accurately characterize a mutant genome relative to a previously sequenced parental (reference) strain. We studied a mutant strain of Pichia stipitis, a yeast capable of converting xylose to ethanol. This unusually efficient mutant strain was developed through repeated rounds of chemical mutagenesis, strain selection, transformation, and genetic manipulation over a period of seven years. We resequenced this strain on three different sequencing platforms. Surprisingly, we found fewer than a dozen mutations in open reading frames. All three sequencing technologies were able to identify each single nucleotide mutation given at least 10-15-fold nominal sequence coverage. Our results show that detecting mutations in evolved and engineered organisms is rapid and cost-effective at the whole-genome level using new sequencing technologies. Identification of specific mutations in strains with altered phenotypes will add insight into specific gene functions and guide further metabolic engineering efforts.

Deletion of BBA64, BBA65, and BBA66 Loci Does Not Alter the Infectivity ofBorrelia burgdorferiin the Murine Model of Lyme Disease

Borrelia burgdorferi, the causative agent of Lyme disease, alters its gene expression in response to highly disparate environmental signals encountered in its tick vector versus vertebrate hosts. Whole-genome transcriptional profile analysis of B. burgdorferi, propagated in vitro under mammalian-host-specific conditions, revealed significant upregulation of several linear plasmid 54 (lp54)-encoded open reading frames (ORFs). Among these ORFs, BBA64, BBA65, and BBA66 have been shown to be upregulated in response to multiple mammalian-host-specific signals. Recently, we determined that there was no significant difference in the ability of BBA64(-) mutant to infect C3H/HeN mice compared to its isogenic control strains, suggesting that B. burgdorferi might utilize multiple, functionally related determinants to establish infection. We further generated BBA65(-) and BBA66(-) single mutants in a noninfectious, lp25(-) clonal isolate of B. burgdorferi strain B31 (ML23) and complemented them with the minimal region of lp25 (BBE22) required for restoring the infectivity. In addition, we generated a BBA64(-) BBA65(-) BBA66(-) triple mutant using an infectious, clonal isolate of B. burgdorferi strain B31 (5A11) that has all of the infection-associated plasmids. There were no significant differences in the ability to isolate viable spirochetes from different tissues of C3H/HeN mice infected via intradermal needle inoculation with either the individual single mutants or the triple mutant compared to their respective isogenic parental strains at days 21 and 62 postinfection. These observations suggest that B. burgdorferi can establish infection in the absence of expression of BBA64, BBA65, and BBA66 in the murine model of Lyme disease.

Burkholderia sabiae sp. nov., isolated from root nodules of Mimosa caesalpiniifolia

Two rhizobial strains, Br3407(T) and Br3405, were isolated from nitrogen-fixing nodules on the roots of Mimosa caesalpiniifolia, a legume tree native to Brazil. On the basis of 16S rRNA gene sequence similarities, both strains were shown previously to belong to the genus Burkholderia. A polyphasic approach, including DNA-DNA hybridizations, pulsed-field gel electrophoresis of whole-genome DNA profiles, whole-cell protein analyses, fatty acid methyl ester analysis and extensive biochemical characterization, was used to clarify the taxonomic position of these strains further; the strains are here classified within a novel species, for which the name Burkholderia sabiae sp. nov. is proposed. The type strain is strain Br3407(T) (=LMG 24235(T) =BCRC 17587(T)).

Multiple pathways for Epstein‐Barr virus episome loss from nasopharyngeal carcinoma

Epstein-Barr virus (EBV) is the prototypical example for episomal persistence of genetic information. Yet, little is known about how this viral episome is lost. Episome loss occurs naturally in nasopharyngeal carcinoma (NPC) upon explantation into culture. Using whole-genome profiling, we found evidence for 2 different pathways of episome loss: (i) rapid loss of the entire episome or (ii) successive mutation/deletion of the episome until at least 1 essential cis-element is destroyed. This second phenotype was seen in a clone of HONE-1 NPC cells that maintains the EBV episome for prolonged time in culture. The conceptual insights provided by our quantitative analysis should aid our understanding of mammalian episomes, as well as lead to designs to cure latent viral infection.

An Integrated Workflow for Analysis of ChIP-Chip Data

Although ChIP-chip is a powerful tool for genome-wide discovery of transcription factor target genes, the steps involving raw data analysis, identification of promoters, and correlation with binding sites are still laborious processes. Therefore, we report an integrated workflow for the analysis of promoter tiling arrays with the Genomatix ChipInspector system. We compare this tool with open-source software packages to identify PU.1 regulated genes in mouse macrophages. Our results suggest that ChipInspector data analysis, comparative genomics for binding site prediction, and pathway/network modeling significantly facilitate and enhance whole-genome promoter profiling to reveal in vivo sites of transcription factor-DNA interactions.

Activation of Notch Signaling in a Xenograft Model of Brain Metastasis

The potential of metastasis can be predicted from clinical features like tumor size, histologic grade, and gene expression patterns. We examined the whole-genome transcriptomic profile of a xenograft model of breast cancer to understand the characteristics of brain metastasis. Variants of the MDA-MB-435 cell were established from experimental brain metastases. The LvBr2 variant was isolated from lesions in a mouse injected in the left ventricle of the heart, and these cells were used for two cycles of injection into the internal carotid artery and selection of brain lesions, resulting in the Br4 variant. To characterize the different metastatic variants, we examined the gene expression profile of MDA-MB-435, LvBr2, and Br4 cells using microarrays. We could identify 2,016 differentially expressed genes in Br4 by using the F test. Various metastasis-related genes and a number of genes related to angiogenesis, migration, tumorigenesis, and cell cycle were differentially expressed by the Br4 cells. Notably, the Notch signaling pathway was activated in Br4, with increased Jag2 mRNA, activated Notch intracellular domain, and Notch intracellular domain/CLS promoter-luciferase activity. Br4 cells were more migratory and invasive than MDA-MB-435 cells in collagen and Matrigel Transwell assays, and the migration and invasion of Br4 cells were significantly inhibited by inactivation of Notch signaling using DAPT, a gamma-secretase inhibitor, and RNA interference-mediated knockdown of Jagged 2 and Notch1. Taken together, these results suggest that we have isolated variants of a human cancer cell line with enhanced brain metastatic properties, and the activation of Notch signaling might play a crucial role in brain metastasis.