Prediction of chromosomal aneuploidy from gene expression data

Significance of Aneuploidy in Predicting Prognosis and Treatment Response of Uveal Melanoma

AIMS

This study aimed to improve personalized treatment strategies and predict survival outcomes for patients with uveal melanoma (UM).

BACKGROUND

Copy number aberrations (CNAs) have been considered as a main feature of metastatic UM.

OBJECTIVE

This study was designed to explore the feasibility of using copy number variation (CNV) in UM classification, prognosis stratification and treatment response.

METHODS

The CNV data in the TCGA-UVM cohort were used to classify the samples. The differentially expressed genes (DEGs) between subtypes were screened by the "Limma" package. The module and hub genes related to aneuploidy score were identified by performing weighted gene co-expression network analysis (WGCNA) on the DEGs. Univariate Cox and least absolute shrinkage and selection operator (LASSO) regression analysis were employed to train the hub genes for developing a prognosis model for UM. Finally, the expression levels of the screened prognostic key genes were verified in UM cells, and the cell migration and invasion abilities were detected using real-time quantitative PCR (qRT-PCR) and transwell assay.

RESULTS

The UM samples were divided into 3 CNV subtypes, which differed significantly in overall survival (OS) and disease-specific survival (DSS). C1 had the shortest OS and DSS and the highest level of immune infiltration. A total of 2036 DEGs were obtained from the three subtypes. Eighty hub genes with the closest correlation with aneuploidy scores were selected by WGCNA. Univariate Cox and LASSO regression-based analyses finally determined eight genes as the key prognostic genes, including HES6, RNASEH2C, NQO1, NUDT14, TTYH3, GJC1, FKBP10, and MRPL24. A prognostic model was developed using the eight genes, demonstrating a strong prediction power. Differences in the response to immunotherapy among patients in different risk groups were significant. We found that high-risk patients were more sensitive to two drugs (Palbociclib_ 1054 and Ribociclib_1632), while low-risk patients were more sensitive to AZD1208_1449, ERK_2440_1713, Mirin_1048, and Selumetinib_1736.

CONCLUSION

UM in this study was divided into three CNV subtypes, and a model based on eight aneuploidy score-related genes was established to evaluate the prognosis and drug treatment efficacy of UM patients. The current results may have the potential to help the clinical decision-making process for UM management.

Hyperdiploidy: the longest known, most prevalent, and most enigmatic form of acute lymphoblastic leukemia in children

Hyperdiploidy is the largest genetic entity B-cell precursor acute lymphoblastic leukemia in children. The diagnostic hallmark of its two variants that will be discussed in detail herein is a chromosome count between 52 and 67, respectively. The classical HD form consists of heterozygous di-, tri-, and tetrasomies, whereas the nonclassical one (usually viewed as "duplicated hyperhaploid") contains only disomies and tetrasomies. Despite their apparently different clinical behavior, we show that these two sub-forms can in principle be produced by the same chromosomal maldistribution mechanism. Moreover, their respective array, gene expression, and mutation patterns also indicate that they are biologically more similar than hitherto appreciated. Even though in-depth analyses of the genomic intricacies of classical HD leukemias are indispensable for the elucidation of the disease process, the ensuing results play at present surprisingly little role in treatment stratification, a fact that can be attributed to the overall good prognoses and low relapse rates of the concerned patients and, consequently, their excellent treatment outcome. Irrespective of this underutilization, however, the detailed genetic characterization of HD leukemias may, especially in planned treatment reduction trials, eventually become important for further treatment stratification, patient management, and the clinical elucidation of outcome data. It should therefore become an integral part of all upcoming treatment studies.

Endoplasmic reticulum-resident protein Sec62 drives colorectal cancer metastasis via MAPK/ATF2/UCA1 axis

OBJECTIVE

Metastasis is responsible for the poor prognosis of patients with colorectal cancer (CRC), and the role of aberrant expression of endoplasmic reticulum (ER) receptors in tumour metastasis has not been fully elucidated. The aim of the study is to ensure the role of ER-resident protein Sec62 in CRC metastasis and illuminate associated molecular mechanisms.

MATERIALS AND METHODS

Bioinformatics analysis, qRT-PCR, western blot and immunohistochemistry assays were performed to evaluate the expression level and clinical significance of Sec62 in CRC. The specific role of Sec62 in CRC was identified by a series of functional experiments. We conducted RNA sequencing and rescue experiments to analyse the differentially expressed genes and identified UCA1 as a novel pro-metastasis target of Sec62 in CRC. Besides, the efficacy of MAPK/JNK inhibitor or agonist on Sec62-mediated CRC metastasis was evaluated by trans-well and wound healing assays. Finally, luciferase reporter and ChIP assay were employed to further explore the potential mechanisms.

RESULTS

The abnormally elevated expression of Sec62 predicted poor prognosis of CRC patients and facilitated malignant metastasis of CRC cells. Mechanistically, Sec62 enhanced UCA1 expression through activating MAPK/JNK signalling pathway. And the p-JNK activating ATF2 could transcriptionally regulate UCA1 expression. Furthermore, blocking or activating MAPK/JNK signalling with JNK inhibitor or agonist potently suppressed or enhanced Sec62 mediated CRC metastatic process.

CONCLUSIONS

Our study reports for the first time that the Sec62/MAPK/ATF2 /UCA1 axis exists in CRC metastatic process, which could be a potential treatment target of metastatic CRC.

Somatic Sex: On the Origin of Neoplasms With Chromosome Counts in Uneven Ploidy Ranges

Stable aneuploid genomes with nonrandom numerical changes in uneven ploidy ranges define distinct subsets of hematologic malignancies and solid tumors. The idea put forward herein suggests that they emerge from interactions between diploid mitotic and G0/G1 cells, which can in a single step produce all combinations of mono-, di-, tri-, tetra- and pentasomic paternal/maternal homologue configurations that define such genomes. A nanotube-mediated influx of interphase cell cytoplasm into mitotic cells would thus be responsible for the critical nondisjunction and segregation errors by physically impeding the proper formation of the cell division machinery, whereas only a complete cell fusion can simultaneously generate pentasomies, uniparental trisomies as well as biclonal hypo- and hyperdiploid cell populations. The term "somatic sex" was devised to accentuate the similarities between germ cell and somatic cell fusions. A somatic cell fusion, in particular, recapitulates many processes that are also instrumental in the formation of an abnormal zygote that involves a diploid oocyte and a haploid sperm, which then may further develop into a digynic triploid embryo. Despite their somehow deceptive differences and consequences, the resemblance of these two routes may go far beyond of what has hitherto been appreciated. Based on the arguments put forward herein, I propose that embryonic malignancies of mesenchymal origin with these particular types of aneuploidies can thus be viewed as the kind of flawed somatic equivalent of a digynic triploid embryo.

SH2B3 inactivation through CN-LOH 12q is uniquely associated with B-cell precursor ALL with iAMP21 or other chromosome 21 gain

In more than 30% of B-cell precursor acute lymphoblastic leukaemia (B-ALL), chromosome 21 sequence is overrepresented through aneuploidy or structural rearrangements, exemplified by intrachromosomal amplification of chromosome 21 (iAMP21). Although frequent, the mechanisms by which these abnormalities promote B-ALL remain obscure. Intriguingly, we found copy number neutral loss of heterozygosity (CN-LOH) of 12q was recurrent in iAMP21-ALL, but never observed in B-ALL without some form of chromosome 21 gain. As a consequence of CN-LOH 12q, mutations or deletions of the adaptor protein, SH2B3, were converted to homozygosity. In patients without CN-LOH 12q, bi-allelic abnormalities of SH2B3 occurred, but only in iAMP21-ALL, giving an overall incidence of 18% in this sub-type. Review of published data confirmed a tight association between overrepresentation of chromosome 21 and both CN-LOH 12q and SH2B3 abnormalities in B-ALL. Despite relatively small patient numbers, preliminary analysis linked 12q abnormalities to poor outcome in iAMP21-ALL (p = 0.03). Homology modelling of a leukaemia-associated SH2 domain mutation and in vitro analysis of patient-derived xenograft cells implicated the JAK/STAT pathway as one likely target for SH2B3 tumour suppressor activity in iAMP21-ALL.

Positive Caricature Transcriptomic Effects Associated with Broad Genomic Aberrations in Colorectal Cancer

We re-examined the correlation between Broad Genomic Aberrations (BGAs) and transcriptomic profiles in Colorectal Cancer (CRC). Two types of BGAs have been examined: Broad Copy-Number Abnormal regions (BCNAs), distinguished in gain- and loss-type, and Copy-Neutral Loss of Heterozygosities (CNLOHs). Transcripts are classified as “OverT” or “UnderT” if overexpressed or underexpressed comparing CRCs bearing a specific BGA to CRCs not bearing it and as “UpT” or “DownT” if upregulated or downregulated in cancer compared to normal tissue. BGA-associated effects were evaluated by changes in the “Chromosomal Distribution Index” (CDI) of different transcript classes. Data show that UpT are more sensitive than DownT to BCNA-associated gene dosage effects. “Over-UpT” genes are upregulated in cancer and further overexpressed by gene dosage, defining the so called “positive caricature transcriptomic effect”. When Over-UpT genes are ranked according to overexpression, top positions are occupied by genes implicated at the functional and therapeutic level in CRC. We show that cancer-upregulated transcripts are sensitive markers of BCNA-induced effects and suggest that analysis of positive caricature transcriptomic effects can provide clues toward the identification of BCNA-associated cancer driver genes.

Single-cell RNA-seq reveals a distinct transcriptome signature of aneuploid hematopoietic cells

Cancer cells frequently exhibit chromosomal abnormalities. Specific cytogenetic aberrations often are predictors of outcome, especially in hematologic neoplasms, such as monosomy 7 in myeloid malignancies. The functional consequences of aneuploidy at the cellular level are difficult to assess because of a lack of convenient markers to distinguish abnormal from diploid cells. We performed single-cell RNA sequencing (scRNA-seq) to study hematopoietic stem and progenitor cells from the bone marrow of 4 healthy donors and 5 patients with bone marrow failure and chromosome gain or loss. In total, transcriptome sequences were obtained from 391 control cells and 588 cells from patients. We characterized normal hematopoiesis as binary differentiation from stem cells to erythroid and myeloid-lymphoid pathways. Aneuploid cells were distinguished from diploid cells in patient samples by computational analyses of read fractions and gene expression of individual chromosomes. We confirmed assignment of aneuploidy to individual cells quantitatively, by copy-number variation, and qualitatively, by loss of heterozygosity. When we projected patients' single cells onto the map of normal hematopoiesis, diverse patterns were observed, broadly reflecting clinical phenotypes. Patients' monosomy 7 cells showed downregulation of genes involved in immune response and DNA damage checkpoint and apoptosis pathways, which may contribute to the clonal expansion of monosomy 7 cells with accumulated gene mutations. scRNA-seq is a powerful technique through which to infer the functional consequences of chromosome gain and loss and explore gene targets for directed therapy.

The acute lymphoblastic leukemia of Down Syndrome - Genetics and pathogenesis

Children with Down Syndrome (DS) are at markedly increased risk for acute lymphoblastic leukemia (ALL). The ALL is of B cell precursor (BCP) phenotype. T-ALL is only rarely diagnosed as well as infant leukemia. Gene expression profiling and cytogenetics suggest that DS-ALL is an heterogeneous disease. More than half of the leukemias are characterized by aberrant expression of the thymic stromal lymphopoietin (TSLP) receptor CRLF2 caused by genomic rearrangements. These rearrangements are often associated with somatic activating mutations in the receptors or in the downstream components of the JAK-STAT pathway. The activation of JAK-STAT pathway suggests that targeted therapy with JAK or downstream inhibitors may be effective for children with DS-ALL. The basis of the increased risk of BCP-ALL and in particular of the CRLF2 aberrations is presently unknown. Neither is it known which genes on the trisomic chromosome 21 are involved.

Consequences of Aneuploidy in Cancer: Transcriptome and Beyond

Cancer cells differ from normal healthy cells in multiple aspects ranging from altered cellular signaling through metabolic changes to aberrant chromosome content, so called aneuploidy. The large-scale changes in copy numbers of chromosomes or large chromosomal regions due to aneuploidy alter significantly the gene expression, as several hundreds of genes are gained or lost. Comparison of quantitative genome, transcriptome and proteome data enables dissection of the molecular causes that underlie the gene expression changes observed in cancer cells and provides a new perspective on the molecular consequences of aneuploidy. Here, we will map to what degree aneuploidy affects the expression of genes located on the affected chromosomes. We will also address the effects of aneuploidy on global gene expression in cancer cells as well as whether and how it may contribute to the physiology of cancer cells.

Haematopoietic development and leukaemia in Down syndrome

Children with constitutional trisomy 21 (cT21, Down Syndrome, DS) are at a higher risk for both myeloid and B-lymphoid leukaemias. The myeloid leukaemias are often preceded by a transient neonatal pre-leukaemic syndrome, Transient Abnormal Myelopoiesis (TAM). TAM is caused by cooperation between cT21 and acquired somatic N-terminal truncating mutations in the key haematopoietic transcription factor GATA1. These mutations, which are not leukaemogenic in the absence of cT21, are found in almost one-third of neonates with DS. Analysis of primary human fetal liver haematopoietic cells and of human embryonic stem cells demonstrates that cT21 itself substantially alters human fetal haematopoietic development. Consequently, many haematopoietic developmental defects are observed in neonates with DS even in the absence of TAM. Although studies in mouse models have suggested a pathogenic role of deregulated expression of several chromosome 21-encoded genes, their role in human leukaemogenesis remains unclear. As cT21 exists in all embryonic cells, the molecular basis of cT21-associated leukaemias probably reflects a complex interaction between deregulated gene expression in haematopoietic cells and the fetal haematopoietic microenvironment in DS.

Genomic instability, driver genes and cell selection: Projections from cancer to stem cells

Cancer cells and stem cells share many traits, including a tendency towards genomic instability. Human cancers exhibit tumor-specific genomic aberrations, which often affect their malignancy and drug response. During their culture propagation, human pluripotent stem cells (hPSCs) also acquire characteristic genomic aberrations, which may have significant impact on their molecular and cellular phenotypes. These aberrations vary in size from single nucleotide alterations to copy number alterations to whole chromosome gains. A prominent challenge in both cancer and stem cell research is to identify "driver aberrations" that confer a selection advantage, and "driver genes" that underlie the recurrence of these aberrations. Following principles that are already well-established in cancer research, candidate driver genes have also been suggested in hPSCs. Experimental validation of the functional role of such candidates can uncover whether these are bona fide driver genes. The identification of driver genes may bring us closer to a mechanistic understanding of the genomic instability of stem cells. Guided by terminologies and methodologies commonly applied in cancer research, such understanding may have important ramifications for both stem cell and cancer biology. This article is part of a Special Issue entitled: Stress as a fundamental theme in cell plasticity.

Analysis of molecular cytogenetic alteration in rhabdomyosarcoma by array comparative genomic hybridization

Rhabdomyosarcoma (RMS) is the most common pediatric soft tissue sarcoma with poor prognosis. The genetic etiology of RMS remains largely unclear underlying its development and progression. To reveal novel genes more precisely and new therapeutic targets associated with RMS, we used high-resolution array comparative genomic hybridization (aCGH) to explore tumor-associated copy number variations (CNVs) and genes in RMS. We confirmed several important genes by quantitative real-time polymerase chain reaction (QRT-PCR). We then performed bioinformatics-based functional enrichment analysis for genes located in the genomic regions with CNVs. In addition, we identified miRNAs located in the corresponding amplification and deletion regions and performed miRNA functional enrichment analysis. aCGH analyses revealed that all RMS showed specific gains and losses. The amplification regions were 12q13.12, 12q13.3, and 12q13.3–q14.1. The deletion regions were 1p21.1, 2q14.1, 5q13.2, 9p12, and 9q12. The recurrent regions with gains were 12q13.3, 12q13.3–q14.1, 12q14.1, and 17q25.1. The recurrent regions with losses were 9p12–p11.2, 10q11.21–q11.22, 14q32.33, 16p11.2, and 22q11.1. The mean mRNA level of GLI1 in RMS was 6.61-fold higher than that in controls (p = 0.0477) by QRT-PCR. Meanwhile, the mean mRNA level of GEFT in RMS samples was 3.92-fold higher than that in controls (p = 0.0354). Bioinformatic analysis showed that genes were enriched in functions such as immunoglobulin domain, induction of apoptosis, and defensin. Proto-oncogene functions were involved in alveolar RMS. miRNAs that located in the amplified regions in RMS tend to be enriched in oncogenic activity (miR-24 and miR-27a). In conclusion, this study identified a number of CNVs in RMS and functional analyses showed enrichment for genes and miRNAs located in these CNVs regions. These findings may potentially help the identification of novel biomarkers and/or drug targets implicated in diagnosis of and targeted therapy for RMS.

An information-theoretic machine learning approach to expression QTL analysis

Expression Quantitative Trait Locus (eQTL) analysis is a powerful tool to study the biological mechanisms linking the genotype with gene expression. Such analyses can identify genomic locations where genotypic variants influence the expression of genes, both in close proximity to the variant (cis-eQTL), and on other chromosomes (trans-eQTL). Many traditional eQTL methods are based on a linear regression model. In this study, we propose a novel method by which to identify eQTL associations with information theory and machine learning approaches. Mutual Information (MI) is used to describe the association between genetic marker and gene expression. MI can detect both linear and non-linear associations. What’s more, it can capture the heterogeneity of the population. Advanced feature selection methods, Maximum Relevance Minimum Redundancy (mRMR) and Incremental Feature Selection (IFS), were applied to optimize the selection of the affected genes by the genetic marker. When we applied our method to a study of apoE-deficient mice, it was found that the cis-acting eQTLs are stronger than trans-acting eQTLs but there are more trans-acting eQTLs than cis-acting eQTLs. We compared our results (mRMR.eQTL) with R/qtl, and MatrixEQTL (modelLINEAR and modelANOVA). In female mice, 67.9% of mRMR.eQTL results can be confirmed by at least two other methods while only 14.4% of R/qtl result can be confirmed by at least two other methods. In male mice, 74.1% of mRMR.eQTL results can be confirmed by at least two other methods while only 18.2% of R/qtl result can be confirmed by at least two other methods. Our methods provide a new way to identify the association between genetic markers and gene expression. Our software is available from supporting information.

Perturbation of fetal hematopoiesis in a mouse model of Down syndrome's transient myeloproliferative disorder

Children with Down syndrome develop a unique congenital clonal megakaryocytic proliferation disorder (transient myeloproliferative disorder [TMD]). It is caused by an expansion of fetal megakaryocyte-erythroid progenitors (MEPs) triggered by trisomy of chromosome 21 and is further enhanced by the somatic acquisition of a mutation in GATA1. These mutations result in the expression of a short-isoform GATA1s lacking the N-terminal domain. To examine the hypothesis that the Hsa21 ETS transcription factor ERG cooperates with GATA1s in this process, we generated double-transgenic mice expressing hERG and Gata1s. We show that increased expression of ERG by itself is sufficient to induce expansion of MEPs in fetal livers. Gata1s expression synergizes with ERG in enhancing the expansion of fetal MEPs and megakaryocytic precursors, resulting in hepatic fibrosis, transient postnatal thrombocytosis, anemia, a gene expression profile that is similar to that of human TMD and progression to progenitor myeloid leukemia by 3 months of age. This ERG/Gata1s transgenic mouse model also uncovers an essential role for the N terminus of Gata1 in erythropoiesis and the antagonistic role of ERG in fetal erythroid differentiation and survival. The human relevance of this finding is underscored by the recent discovery of similar mutations in GATA1 in patients with Diamond-Blackfan anemia.

Classify hyperdiploidy status of multiple myeloma patients using gene expression profiles

Multiple myeloma (MM) is a cancer of antibody-making plasma cells. It frequently harbors alterations in DNA and chromosome copy numbers, and can be divided into two major subtypes, hyperdiploid (HMM) and non-hyperdiploid multiple myeloma (NHMM). The two subtypes have different survival prognosis, possibly due to different but converging paths to oncogenesis. Existing methods for identifying the two subtypes are fluorescence in situ hybridization (FISH) and copy number microarrays, with increased cost and sample requirements. We hypothesize that chromosome alterations have their imprint in gene expression through dosage effect. Using five MM expression datasets that have HMM status measured by FISH and copy number microarrays, we have developed and validated a K-nearest-neighbor method to classify MM into HMM and NHMM based on gene expression profiles. Classification accuracy for test datasets ranges from 0.83 to 0.88. This classification will enable researchers to study differences and commonalities of the two MM subtypes in disease biology and prognosis using expression datasets without need for additional subtype measurements. Our study also supports the advantages of using cancer specific characteristics in feature design and pooling multiple rounds of classification results to improve accuracy. We provide R source code and processed datasets at www.ChengLiLab.org/software.

Recurrent transcriptional clusters in the genome of mouse pluripotent stem cells

A number of studies have shown that transcriptome analysis in terms of chromosomal location can reveal regions of non-random transcriptional activity within the genome. Genomic clusters of differentially expressed genes can identify genomic patterns of structural organization, underlying copy number variations or long-range epigenetic regulation such as X-chromosome inactivation. Here we apply an integrative bioinformatics analysis to a collection of 315 freely available mouse pluripotent stem cell samples to discover transcriptional clusters in the genome. We show that over half of the analysed samples (56.83%) carry whole or partial-chromosome spanning clusters which recur in genomic regions previously implicated in chromosomal imbalances. Strikingly, we found that the presence of such large-clusters is linked to the differential expression of a limited number of genes, common to all samples carrying clusters irrespectively of the chromosome where the cluster is found. We have used these genes to train and test classification models that can predict samples that carry large-scale clusters on any chromosome with over 90% accuracy. Our findings suggest that there is a common downstream activation in these cells that affects a limited number of nodes. We propose that this effect is linked to selective advantage and identify potential driver genes.

Bayesian method to predict individual SNP genotypes from gene expression data

RNA profiling can be used to capture the expression patterns of many genes that are associated with expression quantitative trait loci (eQTLs). Employing published putative cis eQTLs, we developed a Bayesian approach to predict SNP genotypes that is based only on RNA expression data. We show that predicted genotypes can accurately and uniquely identify individuals in large populations. When inferring genotypes from an expression data set using eQTLs of the same tissue type (but from an independent cohort), we were able to resolve 99% of the identities of individuals in the cohort at P(adjusted) ≤ 1 × 10(-5). When eQTLs derived from one tissue were used to predict genotypes using expression data from a different tissue, the identities of 90% of the study subjects could be resolved at P(adjusted) ≤ 1 × 10(-5). We discuss the implications of deriving genotypic information from RNA data deposited in the public domain.

The consequences of chromosomal aneuploidy on the transcriptome of cancer cells

Chromosomal aneuploidies are a defining feature of carcinomas, i.e., tumors of epithelial origin. Such aneuploidies result in tumor specific genomic copy number alterations. The patterns of genomic imbalances are tumor specific, and to a certain extent specific for defined stages of tumor development. Genomic imbalances occur already in premalignant precursor lesions, i.e., before the transition to invasive disease, and their distribution is maintained in metastases, and in cell lines derived from primary tumors. These observations are consistent with the interpretation that tumor specific genomic imbalances are drivers of malignant transformation. Naturally, this precipitates the question of how such imbalances influence the expression of resident genes. A number of laboratories have systematically integrated copy number alterations with gene expression changes in primary tumors and metastases, cell lines, and experimental models of aneuploidy to address the question as to whether genomic imbalances deregulate the expression of one or few key genes, or rather affect the cancer transcriptome more globally. The majority of these studies showed that gene expression levels follow genomic copy number. Therefore, gross genomic copy number changes, including aneuploidies of entire chromosome arms and chromosomes, result in a massive deregulation of the transcriptome of cancer cells. This article is part of a Special Issue entitled: Chromatin in time and space.

DYRK1A in Down syndrome: an oncogene or tumor suppressor?

Children with Down syndrome (DS) have a markedly increased risk of developing acute megakaryoblastic leukemia (AMKL) and acute lymphoblastic leukemia compared with that of children without DS. Despite recent breakthroughs, it is not clear which genes on chromosome 21, the chromosome that is trisomic in individuals with DS, cause this predisposition. In this issue of the JCI, Malinge et al. report their loss- and gain-of-function experiments in mouse and human cells that show that increased expression of the kinase encoded by the chromosome 21 gene DYRK1A suppresses the nuclear factor of activated T cells pathway and promotes AMKL. Interestingly, the same protein has been suggested to contribute to the reduced risk of epithelial cancers in adults with DS, leading to the possibility that it could be proleukemic in children and antitumorigenic in adults.

Virtual CGH: an integrative approach to predict genetic abnormalities from gene expression microarray data applied in lymphoma

Background

Comparative Genomic Hybridization (CGH) is a molecular approach for detecting DNA Copy Number Alterations (CNAs) in tumor, which are among the key causes of tumorigenesis. However in the post-genomic era, most studies in cancer biology have been focusing on Gene Expression Profiling (GEP) but not CGH, and as a result, an enormous amount of GEP data had been accumulated in public databases for a wide variety of tumor types. We exploited this resource of GEP data to define possible recurrent CNAs in tumor. In addition, the CNAs identified by GEP would be more functionally relevant CNAs in the disease pathogenesis since the functional effects of CNAs can be reflected by altered gene expression.

Methods

We proposed a novel computational approach, coined virtual CGH (vCGH), which employs hidden Markov models (HMMs) to predict DNA CNAs from their corresponding GEP data. vCGH was first trained on the paired GEP and CGH data generated from a sufficient number of tumor samples, and then applied to the GEP data of a new tumor sample to predict its CNAs.

Results

Using cross-validation on 190 Diffuse Large B-Cell Lymphomas (DLBCL), vCGH achieved 80% sensitivity, 90% specificity and 90% accuracy for CNA prediction. The majority of the recurrent regions defined by vCGH are concordant with the experimental CGH, including gains of 1q, 2p16-p14, 3q27-q29, 6p25-p21, 7, 11q, 12 and 18q21, and losses of 6q, 8p23-p21, 9p24-p21 and 17p13 in DLBCL. In addition, vCGH predicted some recurrent functional abnormalities which were not observed in CGH, including gains of 1p, 2q and 6q and losses of 1q, 6p and 8q. Among those novel loci, 1q, 6q and 8q were significantly associated with the clinical outcomes in the DLBCL patients (p < 0.05).

Conclusions

We developed a novel computational approach, vCGH, to predict genome-wide genetic abnormalities from GEP data in lymphomas. vCGH can be generally applied to other types of tumors and may significantly enhance the detection of functionally important genetic abnormalities in cancer research.

Identification and classification of chromosomal aberrations in human induced pluripotent stem cells

Because of their somatic cell origin, human induced pluripotent stem cells (HiPSCs) are assumed to carry a normal diploid genome, and adaptive chromosomal aberrations have not been fully evaluated. Here, we analyzed the chromosomal integrity of 66 HiPSC and 38 human embryonic stem cell (HESC) samples from 18 different studies by global gene expression meta-analysis. We report identification of a substantial number of cell lines carrying full and partial chromosomal aberrations, half of which were validated at the DNA level. Several aberrations resulted from culture adaptation, and others are suspected to originate from the parent somatic cell. Our classification revealed a third type of aneuploidy already evident in early passage HiPSCs, suggesting considerable selective pressure during the reprogramming process. The analysis indicated high incidence of chromosome 12 duplications, resulting in significant enrichment for cell cycle-related genes. Such aneuploidy may limit the differentiation capacity and increase the tumorigenicity of HiPSCs.

Amplification of the 20q chromosomal arm occurs early in tumorigenic transformation and may initiate cancer

Duplication of chromosomal arm 20q occurs in prostate, cervical, colon, gastric, bladder, melanoma, pancreas and breast cancer, suggesting that 20q amplification may play a causal role in tumorigenesis. According to an alternative view, chromosomal imbalance is mainly a common side effect of cancer progression. To test whether a specific genomic aberration might serve as a cancer initiating event, we established an in vitro system that models the evolutionary process of early stages of prostate tumor formation; normal prostate cells were immortalized by the over-expression of human telomerase catalytic subunit hTERT, and cultured for 650 days till several transformation hallmarks were observed. Gene expression patterns were measured and chromosomal aberrations were monitored by spectral karyotype analysis at different times. Several chromosomal aberrations, in particular duplication of chromosomal arm 20q, occurred early in the process and were fixed in the cell populations, while other aberrations became extinct shortly after their appearance. A wide range of bioinformatic tools, applied to our data and to data from several cancer databases, revealed that spontaneous 20q amplification can promote cancer initiation. Our computational model suggests that 20q amplification induced deregulation of several specific cancer-related pathways including the MAPK pathway, the p53 pathway and Polycomb group factors. In addition, activation of Myc, AML, B-Catenin and the ETS family transcription factors was identified as an important step in cancer development driven by 20q amplification. Finally we identified 13 "cancer initiating genes", located on 20q13, which were significantly over-expressed in many tumors, with expression levels correlated with tumor grade and outcome suggesting that these genes induce the malignant process upon 20q amplification.

Birt-Hogg-Dubé renal tumors are genetically distinct from other renal neoplasias and are associated with up-regulation of mitochondrial gene expression

Background

Germline mutations in the folliculin (FLCN) gene are associated with the development of Birt-Hogg-Dubé syndrome (BHDS), a disease characterized by papular skin lesions, a high occurrence of spontaneous pneumothorax, and the development of renal neoplasias. The majority of renal tumors that arise in BHDS-affected individuals are histologically similar to sporadic chromophobe renal cell carcinoma (RCC) and sporadic renal oncocytoma. However, most sporadic tumors lack FLCN mutations and the extent to which the BHDS-derived renal tumors share genetic defects associated with the sporadic tumors has not been well studied.

Methods

BHDS individuals were identified symptomatically and FLCN mutations were confirmed by DNA sequencing. Comparative gene expression profiling analyses were carried out on renal tumors isolated from individuals afflicted with BHDS and a panel of sporadic renal tumors of different subtypes using discriminate and clustering approaches. qRT-PCR was used to confirm selected results of the gene expression analyses. We further analyzed differentially expressed genes using gene set enrichment analysis and pathway analysis approaches. Pathway analysis results were confirmed by generation of independent pathway signatures and application to additional datasets.

Results

Renal tumors isolated from individuals with BHDS showed distinct gene expression and cytogenetic characteristics from sporadic renal oncocytoma and chromophobe RCC. The most prominent molecular feature of BHDS-derived kidney tumors was high expression of mitochondria-and oxidative phosphorylation (OXPHOS)-associated genes. This mitochondria expression phenotype was associated with deregulation of the PGC-1α-TFAM signaling axis. Loss of FLCN expression across various tumor types is also associated with increased nuclear mitochondrial gene expression.

Conclusions

Our results support a genetic distinction between BHDS-associated tumors and other renal neoplasias. In addition, deregulation of the PGC-1α-TFAM signaling axis is most pronounced in renal tumors that harbor FLCN mutations and in tumors from other organs that have relatively low expression of FLCN. These results are consistent with the recently discovered interaction between FLCN and AMPK and support a model in which FLCN is a regulator of mitochondrial function.

Comparative gene expression profiling analysis of urothelial carcinoma of the renal pelvis and bladder

Background

Urothelial carcinoma (UC) can arise at any location along the urothelial tract, including the urethra, bladder, ureter, or renal pelvis. Although tumors arising in these various locations have similar morphology, it is unclear whether the gene expression profiles are similar between the upper-tract (ureter and renal pelvis) and lower-tract (bladder and urethra) carcinomas. Because differences may facilitate different screening and treatment modalities, we sought to examine the relationship between urothelial carcinoma of the renal pelvis (rUC) and urothelial carcinoma of the bladder (bUC).

Methods

Fresh tumor tissue was collected from patients with bUC (n = 10) and benign mucosa from the bladder of individuals undergoing resection for non-UC conditions (n = 7). Gene expression profiles from these samples were determined using high-throughput Affymetrix gene expression microarray chips. Bioinformatic approaches were used to compare the gene expression profiles of these samples with those of rUC samples and normal kidney samples that had been described previously.

Results

Using unsupervised analytic approaches, rUC and bUC were indistinguishable. Yet when a supervised analytic approach was used, a small number of differentially expressed genes were identified; these differences were most likely limited to a single pathway--the chloride ion binding activity pathway--which was more frequently activated in rUC than in bUC.

Conclusions

We found that the gene expression profiles of UCs from the upper and lower tract were extremely similar, suggesting that similar pathogenic mechanisms likely function in the development of these tumors. The differential expression of genes in the identified pathway may represent a new avenue for detection of upper-tract tumors.

Hsa-mir-125b-2 is highly expressed in childhood ETV6/RUNX1 (TEL/AML1) leukemias and confers survival advantage to growth inhibitory signals independent of p53

MicroRNAs (miRNAs) regulate the expression of multiple proteins in a dose dependent manner. We hypothesized that increased expression of miRNAs encoded on chromosome 21 (chr 21) contribute to the leukemogenic role of trisomy 21. The levels of chr 21 miRNAs were quantified by qRT-PCR in four types of childhood ALL characterized by either numerical (trisomy or tetrasomy) or structural abnormalities of chr 21. Suprisingly high expression of the hsa-mir-125b-2 cluster, consisting of three miRNAs, was identified in leukemias with the structural ETV6/RUNX1 abnormality and not in ALLs with trisomy 21. Manipulation of ETV6/RUNX1 expression and chromatin immunoprecipitation studies demonstrated that the high expression of the miRNA cluster is an event independent of the ETV6/RUNX1 fusion protein. Overexpression of hsa-mir-125b-2 conferred a survival advantage to Ba/F3 cells following IL-3 withdrawal or a broad spectrum of apoptotic stimuli through inhibition of caspase 3 activation. Conversely, knockdown of the endogenous miR-125b in the ETV6/RUNX1 leukemia cell line REH increased apoptosis after Doxorubicin and Staurosporine treatments. P53 protein levels were not altered by miR-125b. Together these results suggest that the expression of hsa-mir-125b-2 in ETV6/RUNX1 ALL provides survival advantage to growth inhibitory signals in a p53 independent manner.

Genomic imbalances in rhabdomyosarcoma cell lines affect expression of genes frequently altered in primary tumors: an approach to identify candidate genes involved in tumor development

Rhabdomyosarcomas (RMS) are the most common pediatric soft tissue sarcomas. They resemble developing skeletal muscle and are histologically divided into two main subtypes; alveolar and embryonal RMS. Characteristic genomic aberrations, including the PAX3- and PAX7-FOXO1 fusion genes in alveolar cases, have led to increased understanding of their molecular biology. Here, we determined the effect of genomic copy number on gene expression levels through array comparative genomic hybridization (CGH) analysis of 13 RMS cell lines, confirmed by multiplex ligation-dependent probe amplification copy number analyses, combined with their corresponding expression profiles. Genes altered at the transcriptional level by genomic imbalances were identified and the effect on expression was proportional to the level of genomic imbalance. Extrapolating to a public expression profiling dataset for 132 primary RMS identified features common to the cell lines and primary samples and associations with subtypes and fusion gene status. Genes identified such as CDK4 and MYCN are known to be amplified, overexpressed, and involved in RMS tumorigenesis. Of the many genes identified, those with likely functional relevance included CENPF, DTL, MYC, EYA2, and FGFR1. Copy number and expression of FGFR1 was validated in additional primary material and found amplified in 6 out of 196 cases and overexpressed relative to skeletal muscle and myoblasts, with significantly higher expression levels in the embryonal compared with alveolar subtypes. This illustrates the ability to identify genes of potential significance in tumor development through combining genomic and transcriptomic profiles from representative cell lines with publicly available expression profiling data from primary tumors.

Constitutional aneuploidy and cancer predisposition

Constitutional aneuploidies are rare syndromes associated with multiple developmental abnormalities and the alterations in the risk for specific cancers. Acquired somatic chromosomal aneuploidies are the most common genetic aberrations in sporadic cancers. Thus studies of these rare constitutional aneuploidy syndromes are important not only for patient counseling and clinical management, but also for deciphering the mechanisms by which chromosomal aneuploidy affect cancer initiation and progression. Here we review the major constitutional aneuploidy syndromes and suggest some general mechanisms for the associated cancer predisposition.

Association of survival and disease progression with chromosomal instability: a genomic exploration of colorectal cancer

During disease progression the cells that comprise solid malignancies undergo significant changes in gene copy number and chromosome structure. Colorectal cancer provides an excellent model to study this process. To indentify and characterize chromosomal abnormalities in colorectal cancer, we performed a statistical analysis of 299 expression and 130 SNP arrays profiled at different stages of the disease, including normal tissue, adenoma, stages 1-4 adenocarcinoma, and metastasis. We identified broad (> 1/2 chromosomal arm) and focal (< 1/2 chromosomal arm) events. Broad amplifications were noted on chromosomes 7, 8q, 13q, 20, and X and broad deletions on chromosomes 4, 8p, 14q, 15q, 17p, 18, 20p, and 22q. Focal events (gains or losses) were identified in regions containing known cancer pathway genes, such as VEGFA, MYC, MET, FGF6, FGF23, LYN, MMP9, MYBL2, AURKA, UBE2C, and PTEN. Other focal events encompassed potential new candidate tumor suppressors (losses) and oncogenes (gains), including CCDC68, CSMD1, POLR1D, and PMEPA1. From the expression data, we identified genes whose expression levels reflected their copy number changes and used this relationship to impute copy number changes to samples without accompanying SNP data. This analysis provided the statistical power to show that deletions of 8p, 4p, and 15q are associated with survival and disease progression, and that samples with simultaneous deletions in 18q, 8p, 4p, and 15q have a particularly poor prognosis. Annotation analysis reveals that the oxidative phosphorylation pathway shows a strong tendency for decreased expression in the samples characterized by poor prognosis.

Gene set enrichment analysis using linear models and diagnostics

MOTIVATION

Gene-set enrichment analysis (GSEA) can be greatly enhanced by linear model (regression) diagnostic techniques. Diagnostics can be used to identify outlying or influential samples, and also to evaluate model fit and explore model expansion.

RESULTS

We demonstrate this methodology on an adult acute lymphoblastic leukemia (ALL) dataset, using GSEA based on chromosome-band mapping of genes. Individual residuals, grouped or aggregated by chromosomal loci, indicate problematic samples and potential data-entry errors, and help identify hyperdiploidy as a factor playing a key role in expression for this dataset. Subsequent analysis pinpoints suspected DNA copy number abnormalities of specific samples and chromosomes (most prevalent are chromosomes X, 21 and 14), and also reveals significant expression differences between the hyperdiploid and diploid groups on other chromosomes (most prominently 19, 22, 3 and 13)--differences which are apparently not associated with copy number.

AVAILABILITY

Software for the statistical tools demonstrated in this article is available as Bioconductor package GSEAlm.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Mutations of JAK2 in acute lymphoblastic leukaemias associated with Down's syndrome

BACKGROUND

Children with Down's syndrome have a greatly increased risk of acute megakaryoblastic and acute lymphoblastic leukaemias. Acute megakaryoblastic leukaemia in Down's syndrome is characterised by a somatic mutation in GATA1. Constitutive activation of the JAK/STAT (Janus kinase and signal transducer and activator of transcription) pathway occurs in several haematopoietic malignant diseases. We tested the hypothesis that mutations in JAK2 might be a common molecular event in acute lymphoblastic leukaemia associated with Down's syndrome.

METHODS

JAK2 DNA mutational analysis was done on diagnostic bone marrow samples obtained from 88 patients with Down's syndrome-associated acute lymphoblastic leukaemia; and 216 patients with sporadic acute lymphoblastic leukaemia, Down's syndrome-associated acute megakaryoblastic leukaemia, and essential thrombocythaemia. Functional consequences of identified mutations were studied in mouse haematopoietic progenitor cells.

FINDINGS

Somatically acquired JAK2 mutations were identified in 16 (18%) patients with Down's syndrome-associated acute lymphoblastic leukaemia. The only patient with non-Down's syndrome-associated leukaemia but with a JAK2 mutation had an isochromosome 21q. Children with a JAK2 mutation were younger (mean [SE] age 4.5 years [0.86] vs 8.6 years [0.59], p<0.0001) at diagnosis. Five mutant alleles were identified, each affecting a highly conserved arginine residue (R683). These mutations immortalised primary mouse haematopoietic progenitor cells in vitro, and caused constitutive Jak/Stat activation and cytokine-independent growth of BaF3 cells, which was sensitive to pharmacological inhibition with JAK inhibitor I. In modelling studies of the JAK2 pseudokinase domain, R683 was situated in an exposed conserved region separated from the one implicated in myeloproliferative disorders.

INTERPRETATION

A specific genotype-phenotype association exists between the type of somatic mutation within the JAK2 pseudokinase domain and the development of B-lymphoid or myeloid neoplasms. Somatically acquired R683 JAK2 mutations define a distinct acute lymphoblastic leukaemia subgroup that is uniquely associated with trisomy 21. JAK2 inhibitors could be useful for treatment of this leukaemia.

FUNDING

Israel Trade Ministry, Israel Science Ministry, Jewish National Fund UK, Sam Waxman Cancer Research Foundation, Israel Science Foundation, Israel Cancer Association, Curtis Katz, Constantiner Institute for Molecular Genetics, German-Israel Foundation, and European Commission FP6 Integrated Project EUROHEAR.

A comparison study reveals important features of agreement and disagreement between summarized DNA and RNA data obtained from renal cell carcinoma

Cytogenetic abnormalities, such as DNA amplifications and deletions, often lead to significant changes in gene expression levels within a chromosomal region. Instead of generating additional DNA copy number data, one method to identify DNA copy number abnormalities has been to search existing gene expression data for regional perturbations in gene expression. However, it is not clear how well this surrogate method performs in the examination of individual tumors and how we can use both DNA and RNA data to identify candidate genes that may be mutated. Here we report a comparison study using summarized DNA and RNA data to identify chromosomal abnormalities in human samples. Forty-four tissue samples from patients diagnosed as having renal cell carcinoma (RCC) were collected, together with 15 normal kidney samples as controls, and for each sample the genome-wide DNA and RNA data were obtained for comparison using Affymetrix 100K SNP and HGU133plus2 gene expression chips, respectively. The DNA and RNA data was summarized by both chromosome arm and cytogenetic banding patterns and compared. The result of this analysis revealed that the two summarized data sets used to identify cytogenetic changes agreed well. However, some differences between the two were also identified. These differences of large-scale gene expression deregulation without evidence of the comparable DNA copy number alterations may be the result of known mechanisms, such as large-scale methylation or chromosome inactivation, or may be the result of some new mechanism of DNA-RNA translation. The usefulness of the combined data set for identifying regions of mutated genes is also discussed.

An improved method for detecting and delineating genomic regions with altered gene expression in cancer

A method is presented for identifying genomic regions with altered gene expression in gene expression maps.

Genomic regions with altered gene expression are a characteristic feature of cancer cells. We present a novel method for identifying such regions in gene expression maps. This method is based on total variation minimization, a classical signal restoration technique. In systematic evaluations, we show that our method combines top-notch detection performance with an ability to delineate relevant regions without excessive over-segmentation, making it a significant advance over existing methods. Software (Rendersome) is provided.